Skip to main content

Full text of "Electric furnaces in the iron and steel industry"

See other formats










From Advance Sheets of the Second German Edition 
Authorized Translation and Additions 

by C. H. VOM BAUR, E.E. 





Copyright 1913 by 




ELECTRIC furnaces and their use in the manufacture of steel 
and iron have been described in books by Borchers, Neumann 
Askenasy, and others. Their treatises have either described so 
fully the whole subject of electro-metallurgy that only a very 
small space could be allotted to electric iron and steel, or else, 
as in Neumann's volume, only a glance is given at the early 
experiments which were made when these furnaces were first 

Hence, there is need for a work thoroughly describing electric 
furnaces, which are designed only for the steel and iron industry. 

For practical reasons the book is divided into two parts, of 
which the first deals with all questions relative to the construc- 
tion of these electric furnaces, and the apparatus used, while the 
other part takes up the practical use of electric furnaces in the 
steel mill and all its metallurgical reactions. 

While undertaking this work the authors were conscious of 
the difficulty of describing each type of furnace entirely from 
personal observation. This difficulty, however, confronts all 
who are similarly situated, as these electric furnaces have only 
recently been introduced into the iron trades and it is practically 
impossible to know each type from one's own experience. 

As both practical and theoretical men differ regarding the 
advantages of these furnaces for steel and iron making, it is not 
to be expected from this book that any one type of furnace is 
pictured as being better than any other type. Wherever possi- 
ble, therefore, results are given which are based on actual ex- 
perience, although much other material has been used. 






THE preparation of this work in English was undertaken 
in the belief that electric furnaces for the iron and steel industry 
would have their greatest future on the North American Conti- 
nent. Especially is this true of furnaces making electric steel. 
Specifications are daily becoming stricter for steel rails, steel 
castings, and tool steel. Electric steel rails, costing but little 
more than the ordinary kind, are found to be unbreakable in 
service, when laid beside open hearth and Bessemer rails. In 
these latter, scores of breakages have occurred in one season. 
The future of electric steel rails consequently seems assured. 

Electric steel castings have also been on the market for the 
past four years. They are looked upon with favor alike by the 
foundryman and the customer, not only because the highest 
class of steel may be made from the cheapest raw material, but 
also because of the high percentage of good castings and their 
freedom from blow-holes. 

The ability to make homogeneous tool steel, free from gases, 
and at low cost, brought the electric furnace into commercial use 
over a decade ago. In this field it promises to displace com- 
pletely the old and small crucible pot which has been in use since 
the year 1740. 

With these three principal fields now open to electric furnace 
products, it cannot be long before all other domains in the 
use of steel will be invaded. The cost of producing electric steel 
is lower now than that of the crucible process, or of the small 
converter process, and even less than that of the open hearth 
process, as practised with lo-ton furnaces or under. A success 
can, therefore, be confidently predicted for electric furnaces and 
their manufacture of iron and steel. 

A few changes were found necessary, in adapting the German 
to the edition in English, and some fresh material has been added. 



The translator gladly takes this opportunity to thank many 

friends for information and assistance. D^. G. B; Waterhouse, 
of Buffalo, kindly gave the benefit of his extended experience in 
connection with the metallurgy of iron and steel as set forth in 
Part II of this book. To Mr. A. H. Strong, of New York, special 
thanks are due for valuable aid rendered in the various chapters 
on induction furnaces. Mr. Magnus Ugner, of the transformer 
and furnace department of the General Electric Company, very 
kindly read the proofs of many chapters of Part I. No one has 
had a larger or more successful experience in building trans- 
formers and furnaces than Mr. linger. Finally, thanks are due 
to Dr. D. A. Lyon for much new material added, mainly to the 
chapters on electric pig-iron furnaces. 

NEW YORK, September jth t 1912. 


THE realm of Steel and Iron manufacture has in the past ten 
years had a new world of possibilities opened to it by the intro- 
duction of the electric furnace. Before finding a commercial 
foothold among ironmasters, it was in use making ferro alloys. 
Even earlier than this the electric furnace was manufacturing 
aluminum and calcium carbide. 

It has been the aim of the present publication on Electric 
Steel and Iron Furnaces to produce a book for the practical man; 
a comprehensive manual of practical information, yet one ex- 
plaining the electric laws and phenomena involved, and the 
scientific principles upon which the work rests. The under- 
standing of these electrical laws is practically necessary, for in 
electric furnace literature we constantly find assertions con- 
tradicting the simplest of them. The authors also hope in this 
manner to render the book of service to the general student of 
this branch of Electro- Chemical Engineering, and to state 
especially the principal laws which the construction and operation 
of electric furnaces entail, without giving long mathematical 
discussions. Short arithmetical examples nevertheless are given 
dealing with actual furnace problems. Care has been taken to 
mention only those things which have some value in the develop- 
ment of Electric Steel and Iron furnaces, rather than to dwell 
upon theories of little moment. The furnaces most extensively 
used, such as those of Stassano, Heroult, Girod, Kjellin, and 
Rochling-Rodenhauser are described in detail, and compared 
with an ideal Electric Furnace. This seems to be the best 
course to pursue, for in this way an unfair criticism of the differ- 
ent systems can best be avoided. 

In Chapter XIV, "General Review," some furnace designs 
are briefly discussed which have obtained only a limited use or 



which have not yet left the experimental stage, and finally, the 
electric shaft furnace is described at length. 

The discussions are accompanied by a large number of cuts 
and reproductions. 

The demands of actual practise have always been given the 
greatest consideration. Accordingly, the latest results obtained 
from good trials with electrodes in arc furnaces are mentioned, as 
are others of the same order. This volume should, therefore, be 
a welcome adviser to the furnace builder, the student, and in fact 
to anybody who is interested in electric furnaces for the pro- 
duction of steel and iron. 

The authors have written in the hope that these pages will 
aid in the further expansion and success of the electric iron and 
steel trade. 







Some data relating to the development of electrical engineering, ... i 

Tests of Davy and Pepys, 3 

Suggestions by Wall, 4 

by Pichon, 4 

by William von Siemens, 5 

by de Laval, 6 

by Taussig, 8 

The electric furnace of Stassano, 8 

of Heroult, 9 

of Kjellin, 9 

Report of the Canadian Commission under Dr. Haanel, 9 

The electric furnace of Girod, 10 

of Rochling-Rodenhauser, 10 

of Gronwall, Lindblad & Stalhane, 10 


Ohm's Law, 1 1 

Resistance of a conductor, 1 1 

Units of measurement : Ampere, volt, ohm, 12 

Temperature coefficient, 13 

Conductors of the second class, 15 

Series connection, 16 

Parallel connection, 17 

I. Kirchoff's Law, 19 

Combination resistances, 20 

Arithmetical example, 21 

Joule's Law, 23 

The derivation of heat generated, 23 

of power 24 

of work, 25 





The action of heat 26 

1. Direct resistance heating 26 

Gin, Electric furnace of, 27 

Arithmetical example therefore, 28 

Current density, permissible in copper conductors, 29 

2. Induction heating, 32 

3. Indirect resistance heating, 32 

Borchers, Laboratory furnace of, 33 

Hera us, Laboratory furnace of 34 

Girod, Crucible furnace of, 35 

Helberger, Crucible furnace of, 35 

4. Arc heating, 37 

Chemical action, 37 

Motor effect, 39 

Action of two magnets upon each other, 40 

Lines of force of a current-carrying conductor, 41 

Direction of lines of force, 41 

Action between a magnet and electrical conductor, 42 

of two electrical conductors upon each other, 43 

Lines of force of coils, 44 

Pinch effect 44 


Periodicity, frequency, cycle, 47 

Line diagram, 47 

Angular velocity, 48 

Induction, induced currents, 49 

Self-induction, current of self-induction, 50 

Phase difference 51 

Vector diagram, 51 

Coefficient of self-induction, 52 

Apparent resistance, 54 

Power in alternating current circuits, 57 

factor 58 

Losses on account of induction phenomena, 59 

Eddy or Foucoult currents, 59 

Hysteresis losses, 60 

Three phase current, polyphase current, 60 

Star or Y connection, 61 

Delta connection, 62 




Advantages of electric furnaces, 65 

Demands made of an ideal electric furnace, 66 

Influence of the kind of current, 68 

of the frequency, 70 

of changes in the load, 71 

Regulating power of the furnace temperature, 72 

Electric efficiency, 72 

Furnace and hearth arrangement, 73 

Influence of electric heating, 73 

Circulation of the molten metal, 75 

Influence of water cooling, 75 


The arc, 77 

Radiation furnaces, 79 

Combined arc and resistance furnaces, 79 

The electrodes of arc furnaces, 80 

Current density in electrodes, 82 

Efficiency of electrodes, 84 

Burning away of electrodes, 89 

Electrode consumption, 89 

coverings, 97 

cooling, 99 

regulation, 102 

Thury, regulator, 102 


Stassano shaft furnace, 107 

hearth furnace, 108 

rotating furnace, 108 

Comparison with an ideal furnace, . . . . % 116 

Installation costs, 120 

Issuing of licenses, 120 



Historical, 121 

The Furnace, 122 

Comparison with an ideal furnace, 133 

Installation costs 141 

Issuing of licenses, 143 




Historical, 144 

The furnace 145 

Comparison with an ideal furnace, 151 

of electrode cross-section with a Girod and Heroult, . . . .154 

Installation costs, 157 

Issuing of licenses, 159 


Principle of the transformer, 160 

of the induction furnaces 161 

Cylinder winding, tube winding, disk winding, 165 

Suggestions by de Ferranti, 165 

by Colby, 169 

by Kjellin, 169 

by Frick, 169 

Arrangement for lessening the stray fields, 171 

Suggestion by Rochling and Rodenhauser 172 



Historical, 173 

The furnace, 173 

Influence of the furnace contents on the power factor, 178 

Comparison with an ideal furnace, 185 

Issuing of licenses, 192 



Its beginning, 193 

The furnace, 197 

Regulating transformers, auto transformers, 214 

Installation costs, 223 

Issuing of licenses, 224 


The Stassano electric shaft furnace, 225 

The Keller electric shaft furnace, 226 

The Heroult electric shaft furnace, 227 



The test furnaces of Gronwall, Lindblad & Stalhane 228 

The Gronwall, Lindblad & Stalhane electric shaft furnace, 231 

Influence of carbon on the energy taken up, 235 

Results of operation, 235 

Installation costs, 239 

Issuing of licenses, 240 

Statistics 241 


The Chapelet arc furnace (Giffre, Allevard), 242 

The Keller arc furnace, 244 

The Nathusius arc furnace, 245 

The Gin induction furnace, 248 

The Schneider-Creusot induction furnace, 248 

The Gronwall, Lindblad & Stalhane induction and arc furnace, . . . 249 

The Hiorth combination furnace and induction furnace, 250 

The Baily heating furnace, 253 


Economical, 257 

Statistics, 261 




Their general requirements, 278 

"Schamotte" fire-bricks, . , 280 

Acid or silica bricks, 281 

" Half Schamotte " fire-bricks, - 281 

Carbon bricks and carbon for ramming in place, 281 

Basic bricks and materials for ramming in place, 282 

Chrome iron ore, 282 

Dolomite, 282 

Dolomite plant, 282 

Tar, 282 

Magnesite and magnesite bricks, 283 

Mortar, 283 

Fluxes for the rammed part of the lining, 284,, 

Form of hearth and durability of lining, 286 " 




Influence of the kind of furnace on the quality of steel 286 

General operating costs, 287 

Charge, 287 

Loss in working, 288 

Comparison of the heating cost in the open-hearth and electric furnace, 289 
Comparison of the heating cost in the crucible and electric furnace, . .291 

Statistics concerning electric steel production in Austria, 293 

The amount of power used and its influence 294 

Comparison of the heating cost in the electric shaft and ordinary blast 

furnace 295 

Unit price for electric power, 296 

Slag-making materials, 297 

Labor, 297 

Costs for lining or furnace maintenance, 298^ 

Amortization costs, 299 

Cost of electrodes, 300 

Auxiliary arrangements, 300 

Consumption of tools 301 

Total costs of operation of the electric shaft compared with the ordinary 

blast furnace, 302 

Total costs of operation of the Stassano furnace, * ,. 304 

Total costs of operation of the Girod furnace, 306 

Total costs of operation of the Heroult furnace, 307 

Increased cost through desulphurization by means of Ferro-Silicon, . . 307 

Total costs of operation for the Rochling-Rodenhauser furnace, . . . 308- 


Introduction, 310 


The smelting of ore in the Stassano furnace, 319 

The smelting of ore in the Gronwall, Lindblad & Stalhane electric 

shaft furnace, 322 

The smelting of ore in the Rochling-Rodenhauser induction furnace, . . 323 

Chemical balance, 325 

Smelting results, 327 

Criticism of ore smelting in the electrode-hearth furnace, 333 

The smelting of ore in the electrode shaft furnace, 335 

Ore smelting tests in the special Heroult furnace, 339 

Criticism of this method of smelting, 344 

Ore smelting in the Gronwall, Lindblad & Stalhane furnace, . . . 345 

Efficiency of the furnace, 352-356 

Crtiicism of the furnace, 360 





The impurities in steel: phosphorus, sulphur, silicon, copper, arsenic, 

carbon, oxygen, manganese, aluminum, vanadium, titanium, . . . 365 

The slag-producing materials, ferro alloys, etc., used in the electric fur- 
nace: Ferro-manganese, ferro-chrome, ferro-silicon, lime, fluor-spar, 

iron ore, carbon, 376 

The electric furnace as a melting furnace for iron and steel and iron 

alloys of every kind, 378 

Melting of pig iron, 379 

Melting of ferro-manganese, 380 

The electric furnace as a mixer, 382 

Pig-iron refining, 383 

Production of special quality steel in the electric furnace, 387 

From previously refined metai with low phosphorus and sulphur, . 388 
From previously refined metal with considerable phosphorus and 

sulphur, 392 

The metallurgical course of operations of an electric furnace charge, . . 397 


Final considerations, 400 

Comparison of heating costs in the open-hearth and electric furnace, . . 401 
Index 405 



A = Work or Energy. 
Cos (f) = Power factor. 

E = Potential per phase. 

e = Potential in volts = Effective value for A. C. 
e = Maximum value of potential. 
e' = Instantaneous value of potential. 
CL = Potential to overcome the self-induction. 
e r = Potential to overcome the ohmic resistance. 
I = Current per phase. 

i = Current in amperes = Effective value for A. C. 
i r = Watt component of current. 
i m = wattless component of current. 

k = Heat conductivity. 
KVA = Kilo volt- amperes. 
KW = Kilowatt. 
K\V Hr = Kilowatt hours. 
L = Self-induction. 
1 = Length of a conductor in metres, 
m = Angular velocity. 
N = Flux, 
p = Power. 

p' = Instantaneous value of power. 
Q = Energy in heat, units. 
q = Section in square millimetres, 
r = Resistance in ohms. 
s = Turns. 
T = Time of a cycle, 
t = Time. 
VA = Volt amperes. 

A = = Current density per square millimetre. 


v = Cycles per second. 

p = Specific resistance per I metre length and I square millimetre 

= Specific conductivity. 

pi = Specific resistance per cubic centimetre. 
PI = Specific resistance per cubic inch. 

= x = Specific conductivity per centimetre. 

r = Temperature gradient. 

d = Diameter. 



Part One 



GREAT interest is today manifested in electric steel and its 
production. Not only are the different iron and steel works 
installing electric furnaces or considering their adoption when 
enlargements become a good investment, but political economists 
are also carefully following the progress made in the electric- 
steel industry. The daily press frequently contains accounts of 
the importance of electric iron and steel. Considering the great 
and almost universal interest shown today in the new industry, it 
must seem astonishing that but ten years ago hardly a thought 
was given to the practical utilization of the electric furnace for 
producing steel. This remarkable growth originating in the 
laboratory seems to justify us in following the development of 
the electric furnace, and in tracing the causes which have made 
its entrance into the great industries possible. 

In the first instance we must clearly understand that all 
electric furnaces are naturally apparatus in which electrical 
energy is consumed for the purpose of transforming it into heat. 
Their development on a large scale was therefore not possible 
until electrical engineering had succeeded in producing sources 
of current which furnished it economically, continuously and of 
sufficient size. 

At the beginning of the last century thermopiles, or galvanic 

cells, as they are used today for operating house bells, or telephone 



circuits, were the only sources of electric current at the disposal 
of the user of electricity, so that we see the electric current of 
this period confined in its application to the laboratories of the 
scientist. It was only toward the middle of the igth century 
that a strong development started which has its foundation in 
Faraday's discovery of induction. 

In 1831 Faraday found that each time he brought a strong 
magnet near to, or moved it away from, a coil of wire, the ends of 
which were separated a very small distance from each other, a 
tiny spark appeared at the point of interruption. We say, 
therefore, that Faraday found that the magnet induces a current 
in the electric conductor (the coil of wire). 

This discovery brought a great light into the darkness which 
until then had covered the practical generation of electricity; 
for, hardly a year after Faraday's discovery, we find the first 
magnet-electric machine, which was built by Pixii. This was 
the first form of a dynamo machine, that is, of a machine which 
transforms rotary motion into electric energy. In Pixii's ma- 
chine a coil of wire was arranged in the magnetic field of a strong 
ordinary horseshoe magnet, in such a way that when the coil 
w r as rotated, induced currents were produced, as discovered by 
Faraday. This first machine was soon followed by improved 
designs, which, even at this early period of electrical science in 
the fifties of the last century, were put to use in supplying light 
in lighthouses on the coasts of France and England. 

The next step forward was accomplished by H. Wilde in 1866 
in Manchester, by the construction of an electric machine, whose 
magnets were electro-magnets. For these a small machine 
with ordinary magnets furnished the current. 

A Wilde generator of this type, which required about 3 h. p. 
for driving the exciting machine and about 15 h. p. for the main 
dynamo, was able to melt a bar of platinum 6 mm. thick (about 
y^ inch) and 60 cm. (two feet) long. 

The above mentioned electric magnetic machines were, 
however, despite their considerable power, unable to introduce 
electricity for general uses. An important forward step was 
still lacking until Werner von Siemens discovered the " dynamo 


electric principle," which he laid before the Berlin Academy on 
January 17, 1867. According to this principle the "residual" 
magnetism that remains in even the softest iron is sufficient to 
produce an extremely weak current by which the magnetism 
can be strengthened more and more. The employment of this 
discovery in the construction of dynamos now made it possible 
to use electro-magnets instead of the ordinary permanent magnets 
heretofore used, and, with this improvement, we have the dynamo 
as it is today. It is used the world over, following the Siemens 

But even after Siemens' discovery considerable time elapsed 
before any great change occurred in the output of large electric 
generators. It was the invention of the incandescent lamp, 
first generally known in Europe through the Paris international 
electric exhibition in 1881, which brought about this develop- 
ment. Electrical power-houses in ever increasing numbers and 
sizes now appeared, and today we see them in nearly every city. 
If we finally call to mind the well-known first great power trans- 
mission of a current at 30,000 volts pressure over a distance of 
170 km. (106 miles), between Lauffen and Frankfurt, Germany, 
which was shown to the world in 1891 at the time of the Frank- 
furt Exhibition, we find ourselves in the midst of tremendous 
advances of electrical science. 

If we now turn to the history of the development of the 
electric furnace itself, we find its first traces at the beginning of 
the i gth century; that is, at the same time in which the sole 
means of producing electric currents was the thermopile, and at 
which time no thought of any electrical science existed. The 
first to consider the practical exploitation of electric energy by 
converting it into heat, probably was Davy, who, about 1810, 
during his experiments in the electrolysis of aluminum oxide, ex- 
cluded any influx of heat from the outside, and produced the 
heat necessary for the experiment by the electric current itself, 
which he obtained from an apparatus naturally very inferior, 
from the view-point of our modern ideas. His apparatus consist- 
ed of a platinum plate connected with one pole of a thermopile 
of 1000 plates, the other pole being connected with an iron wire. 


The latter projected from the upper side into a layer of clay 
carried by the platinum plate, which was in connection with the 
other pole. When the circuit was established the iron wire 
became white hot and melted where it was in contact with the 

A far more perfected arrangement was that of Pepys, who 
in 1815 welded an iron wire by heating it with an electric current. 
Pepys' apparatus can be looked upon as the first form of the 

FIG. i. 

class of electric furnaces today known as resistance furnaces. A 
soft iron rod was slotted with a fine saw in the direction of its 
axis and the slot was filled with diamond dust. The rod was 
then wound with wire and heated to red heat for 6 minutes by 
means of an electric battery (Fig. i). An examination of the 
iron wire showed that the diamond dust had disappeared, and 
that the iron had changed with the absorption of carbon. In 
this experiment iron was for the first time treated by the applica- 
tion of electric heat. 

It is interesting to find that in 1843 A. Wall made the suggest- 
ion that pig iron be treated and converted by electrical means. 
In 1853, I0 years later, we find in a French patent granted to 
Pichon, the first electro-thermic furnace. The patent claim is 
as follows: " economical and application of the electric light to 
metallurgy and particularly metallurgy 
of iron." The furnace reproduced by 
Fig. 2 shows the original design of a 
furnace indirectly heated by electric 
arcs. Such furnaces are used even to- 
day with some changes in the design 
given us by Stassano. The ore or 
metal which Pichon tried to melt in his 

furnace was dropped between electrodes of considerable area 
through which the electric current passed. It was expected 


that the charge would melt under the influence of the tempera- 
ture of the arc, and collect in the bottom, which in turn was to 
be heated. Pichon's idea was to build his type of furnace on a 
large scale, this being clearly indicated by the dimensions of the 
electrodes which were to be 3 m. (10 ft.) long and to have a 
cross-section of 60 sq. cm. (9.3 sq. inches). 

It is interesting to observe that Pichon's suggestion appeared 
exactly at the time in which the first attempts were being made 
to illuminate the sea-coasts by means of electric light supplied 
by permanent magnet-electric dynamos. Electrical science, 
which thus called this furnace into existence, was, however, 
unable to further the realization of Pichon's daring plans, so 
capable of life as later developments show. The magnet-electric 
machine was by a large margin incapable of furnishing the current 
necessary for the operation of Pichon's furnace. 

Many different schemes were tried, in the years immediately 
following Pichon, to utilize the electric current in the production 
of iron, but they failed, being in advance of their time. The 
English patents of William von Siemens, of the years 1878 and 
1879, next bring developments in the design of electric furnaces. 
They contain nearly all the important details of the modern arc 
furnaces, and for this reason they will be examined somewhat 
more closely. Siemens used different types of furnaces. The 
first design consisted of a crucible surrounded by a 
metallic case, through the bottom of which pro- 
jected one pole of an electric circuit. That part 
of the electrode in direct touch with the charge 
was provided with a point of platinum or .other 
substance capable of resistance of great heat, in 
order to avoid contaminating the charge. The 
second electrode, which was connected with the 
other pole of the electric circuit, entered through 
the cover of the furnace and was cooled with 
water or other liquid. Figure 3 shows the arrangement of the 
furnace. Siemens later changed the design, making the elec- 
trode, which entered from the top, of carbon, while the lower 
metallic electrode was cooled with water. A heat-protecting 


FlG. 4. 

covering was provided for this furnace. The crucible was 
placed in a larger case of metal and the space between the 
two filled with charcoal or other poor conductor of heat, as 
shown in the design, Fig. 4. 

Siemens finally used a form of furnace 
very similar to that of Pichon. Two carbon 
electrodes were inserted in the sides of a 
crucible in such a position that they remained 
above the top of the charge, and the arc 
formed between them did not come in con- 
tact with the material to be melted. This 
furnace is shown in Fig. 5. With it Siemens 
succeeded in melting 10 kg. (22 Ibs.) of 
steel per hour. He also reduced iron ore 
and fused metals of high melting point such as platinum, taking 
about one-quarter of an hour to liquefy 4 kg. (8.8 Ibs.) of 
the latter substance. Siemens figured theoretically that the 
combustion of i kg. of coal under the boilers of a dynamo- 
electric generating plant would produce i kg. of melted steel. 

Siemens' furnaces, in regard to their practical construction, 
attained a high degree of perfection. They were equipped with 
automatic devices for adjusting the carbons and to keep the arc 
length always the same. He also utilized the directive qualities 
of the electro-magnet in order to obtain 
the best heating effects. All modern 
constructions of arc furnaces are adapta- 
tions of this original design, differing 
simply in size and form and other minor 
respects. The reason why the Siemens 
furnace failed of successful introduction on a commercial scale, 
lies in the fact that current was still too expensive; it cost 
too much in those days to be of use in melting iron in electric 

With the suggestions of Siemens, the furnace subject seemed 
for the time exhausted. Aside from a long list of unimportant 
patents the ensuing time shows no progress until the appearance 
of the interesting patent of de Laval, of the year 1892. Fig. 6 

FIG. 5. 


shows this furnace. The hearth of a cylindrical furnace is 
divided into two parts by a bridge cooled with water. At the 
bottom of each of the two compartments metal or carbon elec- 
trodes are inserted and connected with a source of alternating 

FIG. 7. 

FIG. 8. 

current. The furnace was to be charged from the top by re- 
moving the cover and first introducing a quantity of molten 
magnetic oxide of iron. The succeeding charges were to be of 
spongy iron. 

In operating the furnace it was intended to cover the bridge 
with a layer of oxide or other fusible substance which would act 
as a resistance. With an iron-refining furnace oxidized iron was 
to be used. The object sought was to have the spongy iron 
undergo a refining process in falling through the material forming 
the resistance; and this was the purpose for which de Laval's 
furnace was designed. De Laval saw a great future for this 
furnace and the extent of his hopes can best be realized from the 
fact that with Nobel, in 1895, he laid plans 'for a power plant of 
some 35,000 h. p., to be used in melting iron by electricity. 
These bright hopes failed of realization, and de Laval's furnace 
has today an historical interest only, as the first example of a 
furnace to melt iron by direct resistance. But the plans show 
that we have arrived at an epoch in the development of electro- 
chemistry, by aid of which, and that of great water-power it is 
possible to consider operating electric furnaces on a large scale. 


Another type of resistance furnace, the Taussig, appeared in 
1893. No longer is a separate liquid resistance required. The 
charge itself, whether metal or ore, forms the resistance, and in 
consequence the furnace takes the shape of a horizontal groove 
rather than a vertical one as Fig. 7 shows. This furnace also 
remained unused. 

It is evident from the above that in the middle of the nineties 
of the last century there existed a number of designs for electric 
furnaces, and that the process of heating metal baths by elec- 
tricity was well understood. But the iron industry had so far 
refused to take the electric furnace seriously. This is even more 
astonishing considering the pathfinding and successful employ- 
ment of the Heroult furnaces of 1887 and 1888 in the aluminum 
industry, and the use of the electric furnace in the manufacture 
of calcium carbide in 1894. Both these industries had already 
attained their full growth when, with the new century at last, 
the interest of the iron industry in the electric furnace began to 
awaken. The main reasons for this late beginning of the electro- 
steel and electro-iron industries may be sought, first, in the 
high development of the existing process for the manufacture 
of steel and iron, which seemed to preclude any possibility 
of cheapening steel; and, second, in the fact that nothing 
definite was known about the quality of the product of the 
electric furnace. 

The first practical constructions of furnaces to melt and 
refine iron appeared at the same places where the iron industry 
itself had come into being; that is, places having favorable 
water-power. Here the electric furnaces could obtain cheap 
current for experimental purposes. This leads to that 
point of the development which produced the present furnaces, 
to be considered more closely in later chapters; at this time, 
therefore, only the historical facts will be recorded in a general 

In 1898 Stassano took out a patent in different countries 
claiming: "A method for the practical production of liquid 
wrought iron of any degree of carbon and of liquid alloys of iron 
by means of the electric current." Stassano's furnace under- 


went many constructive alterations as a result of experiments 
made to obtain a practical apparatus, but his furnaces even as 
used today are based on the old principle of heating. 

The next most important type of furnace used today, the 
Heroult, appeared in the years 1899 and 1900, and almost at the 
same time Kjellin with his induction furnace succeeded in pro- 
ducing an apparatus of practical use in the iron industry. 

All these three furnaces were operated by electricity generated 
by means of water-power. The Stassano in upper Italy, the 
Heroult in Savoy, and the Kjellin in Sweden, and their practical 
success, first drew the interest of the iron industry. An impor- 
tant contributing factor also was a report by Dr. Haanel, chief 
of a commission of experts sent by the Canadian Government to 
Europe to study the electric furnace. 

This report first brought together the different existing 
types of furnaces and considered them in their relation to each 
other. The plants of Gysinge, Kortfors, La Praz, Turin, and 
Livet were inspected, and it was found that a Kjellin furnace in 
Gysinge produced a superior quality of steel from raw materials 
consisting of charcoal iron and scrap iron. In Kortfors, and 
also in La Praz, the Heroult steel process was in operation, any 
desired quality of steel being produced by the method of first 
melting scrap and then refining it by the use of a large variety 
of slags. The Stassano furnace in Turin was not in operation 
at the time of the commission's inspection. In Livet a furnace 
by Keller, of a construction similar to that of Heroult, was busy 
melting iron direct from the ore. 

From the above it is evident that in 1904, the time of the 
Canadian Commission's tour, the electro-steel industry had at- 
tained a healthy existence, at least where in proximity to water- 
power developments. The principal hindrance to the introduc- 
tion of the electric furnace in the iron industry had now been 
overcome. In the production of steel the problem was to keep 
the iron from absorbing the carbon of the electrodes, and both 
Kjellin and Heroult successfully solved the difficulty although in 
different ways. Kjellin avoided the use of electrodes entirely, 
while Heroult brought the electric current into the furnace 


through carbon electrodes, following the method used (for ex- 
ample) in the aluminum industry, where the electrolysis of the 
molten mass is desired. He, however, arranged his furnace so 
that there always remained a layer of slag interposed between the 
metal and the carbon, thus avoiding contact between the two. 

Stassano sought to attain his goal in the reduction of iron 
directly from the ore and only later turned to the scrap-iron 
method. Heroult and Kjellin were the first to regularly engage 
in the business of melting scrap iron in the electric furnace. 
And again it is Heroult to whom credit is due for the develop- 
ment of the art, to the end that from a charge of ordinary scrap 
any desired quality of steel may be obtained by refining it. 

Improvements in the machinery for generating electric 
currents, especially in the design of gas engines of large capacity, 
had in the mean time opened the way for other cheap methods of 
producing electricity, so that the electro-steel industry was no 
longer limited to water-power locations. In 1905 there appeared 
the first such industry in Germany, in the works of Richard 
Lindenberg in Remscheid. The installation consisted in one 
Heroult furnace. 

In the same year the Rochling Iron and Steel Works erected 
a Kjellin furnace and were the first to try the experiment of 
running an electric furnace in conjunction with a great iron 

There remains to be mentioned that during this period the 
now much used Girod furnace appeared in a small way, and that 
the year 1906 was marked by the appearance of the Rochling- 
Rodenhauser furnace, of which the following chapters will 
speak more fully. With the latter there now existed an induction 
furnace by means of which (as also with the Heroult furnace) 
a superior steel of any desired quality could be obtained from a 
charge of any kind of raw material. 

In recent years successful efforts have been made to produce 
iron by means of the electric-shaft furnace of Gronwall, Lindblad 
& Stalhane. More modern times have arrived and in the following 
chapters will be found a discussion of the present day furnaces 
as used in the electric-steel industry. 



BEFORE turning to a discussion of the different types of 
electric furnaces used today, it is necessary to have a clear under- 
standing of some of the fundamental electrical laws and terms, 
for it is only through a knowledge of this, that an electric furnace 
can rightly be judged and a correct picture conceived of the 
occurring phenomena. In order to begin with the most impor- 
tant foundation for all electrical investigations we have first to 
deal with Ohm's law. This law says: 

Drop in potential 
Current = - JT . f 


or if i denotes the current, e the drop in volts and r the resistance 

The resistance r of a conductor is determined by the equation : 



where / denotes the length and q the cross-section of the con- 
ductor, while c is a constant depending upon the material. 

It can be shown, for instance, that under the same electrical 
conditions and measurements, a copper conductor will convey 
5^ times the current that an iron conductor will. The reason 
underlying this is that evidently copper conducts electricity 
better than iron. We therefore speak of the different conduc- 
tivities of different materials. 

The above-mentioned phenomena may, however, be explained, 
since different materials give entirely different resistances, even 
though the dimensions may be the same. The resistance factor 
dependent on this material is called the "specific resistance" 



and is mathematically indicated by p, so that the formula for 
the resistance of a conductor is, 

/ / 

r = c and may now be written : r = p 

9 q 

The different conductivities are consequently the reciprocal 
values of the specific resistances. 

In order to be able to apply our first law, the "ohrnic law/' 
we must clearly establish the electrical units. The current 
quantity is measured in Amperes, the potential or pressure in 
Volts, and the resistance in Ohms. 

Originally the unit of resistance as established by Siemens 
consisted of a column of mercury one metre long and of one 
square millimetre cross-section at a temperature of o centigrade. 
In place of this resistance unit, we have the ohm to-day, which 
equals 1.063 Siemens units and corresponds to a mercury column 
1.063 m - l n g> f z sc i- mm - cross-section at o C. 

The generally applied resistance unit, the ohm, was established 
in such a way that, with a pressure of one volt, and a resistance 
of one ohm, a current of one ampere resulted. 

We now know what the dimensions of the mercury column 
are, and under which conditions it has a resistance of one ohm, 
and since the resistance of a conductor is, 

/ / is in metres 

r = p where 

q q in square millimetres 

we may calculate the specific resistance for mercury, which is 

1.063 m - 

i ohm = p - or p = i : 1.063 == .94073 

i sq. mm. 

In the above description of the resistance unit the temperature 
was always given. This was not done without having an object 
in view, for the specific resistance of a conductor is not only 
dependent on the material, but also on its temperature. There- 
fore, equal conductors at different temperatures have different 
resistances, and consequently with the same voltage they carry 
different currents. 


The conditions governing the alterations of the specific 
resistance with changing temperatures have been established, 
by making exact measurements with the different materials. In 
a similar way the specific resistances were determined. 

Concerning the changes in temperature, it was shown 
that the resistance of the metals and their alloys rose with 
increasing temperatures and in accordance with the following 

r t = r (i + a t -f f) 

where r = the resistance at o 

r t = the resistance at f 

a and |8 are numerical constants, which have to be specially 
determined for each conductor. 

For practical purposes when within moderate temperature 
differences, it will suffice if we use the following formula: 

r t = r (i -f a t) 

It is well known that the specific resistance of a metal de- 
pends so much on various substances mixed with it, that accurate 
figures, as they are known for the absolutely pure metal, have 
practically only a comparatively small value. It is sufficient 
therefore to figure practically with the following values: 

Material p 

Copper %s 

Brass, Iron, Platinum, Zinc Vio 

German Silver, and similar German Silver alloys: 

Monel metal % to ]/2 

Carbon varies between '. . . . 100 and 1,000 

The exact values of the specific resistances, and the tempera- 
ture coefficients, are given for some of the materials which are 
used in the construction of electric furnaces. It is to be noted 
that the temperature coefficient for metals is positive, i.e., with 
rising temperature its resistance increases. In contradistinc- 
tion to this, carbon and non-metallic conductors are negative, i.e., 
with rising temperatures the resistance is decreased. 



. P at 15 C 



Q-I to OS 

+ OO^Q 



-j- OO4I 


IO tO 12 

+ OO4.S 

German Silver 

.018 to .019 
.07 to .08 
IS to i>6 

+ .0037 
+ .0015 
-j- OOO2 tO 



1 5 

+ -OO37 


.12 to . 16 
.016 to .018 

+ . 0024 to 
-f- . 0034 to 




. 10 to .25 

+ . 0052 



+ . 0042 


100 to 1000 

. 0003 to 


The temperature coefficients in the above table are only 
exactly accurate within comparatively small temperature in- 
tervals, in fact, are absolutely accurate only within the limits 
of o to 30 C. This is why the temperature coefficient only 
gives an approximate idea of the increase in the specific resist- 
ance. For instance, that which interests us the most is iron, 
with its growing temperatures. Unfortunately exact measure- 
ments of the resistance of iron at high temperatures are extremely 
difficult to make, and this is why we see so many contradictory 
statements concerning these. It is evident that the resistance 
of solid and also fluid iron varies with its chemical composition. 
To some extent a certain portion of the iron content will include 
gases and slag particles at the beginning of the run, and this 
causes it to have a higher resistance than it would have at the 
end of the heat, yet keeping the composition the same. But all 
of these influences are practically negligible; for in the many 
years' experience of the author in operating electric furnaces, 
there has never been any significant or practically real influence 
of the refining which could be credited to a change in the re- 
sistance of the iron. Thus the above-mentioned causes, which 
could theoretically call forth a change in the resistance, may be 
neglected in practise. We may now use the formulas and values 
which are sufficiently correct for practical purposes, and which 
are obtained by exact measurements for low temperatures. It 


has been established that the specific resistance of ordinary basic 
bessemer iron at temperatures from o to 160 C., would change 
according to the following formulas: 

Pt = .13 (i + 5 X icT' 5 / + 3.6 X iQ- 6 / 2 ) 

If in accordance with this formula we figure the value of p t for 
1700, we obtain 

piyoo = 2.12 

This result seems somewhat too high, as far as it can be 
judged with the operating values on hand which were obtained 
by the author, who has had the best results when figuring with 
a mean value of p = 1.66 when designing electric furnaces. 
GIN figured according to Neumann with a resistance of iron 
of .000175 ohm per cubic centimetre. This would correspond 
to a specific resistance of p = 1.75. This figure also shows that 
the result calculated above for the specific resistance of fluid 
iron is too high at 2.12. In this book p = 1.66 will always be 
used as the specific resistance of molten iron. Even though 
this value cannot lay claim to any theoretical accuracy, the 
calculations will, however, give results which correspond 
sufficiently with practical operating conditions. 

It was remarked upon before that carbon, in contradistinc- 
tion to the metals, has a negative temperature coefficient, so 
that with increasing temperatures the resistance of the carbon 
decreases. This phenomenon we have, however, in a much 
more extraordinary measure, with the so-called " conductors of 
the second class." Under this heading we mean materials, or 
bodies, which at ordinary temperatures have practically no 
conductivity, or at least so small a one as not to be worthy of 
consideration. With increasing temperatures these conductors 
of the second class attain steadily bettering conductivities, so 
that they can eventually be used as conductors directly. We 
shall have to deal with conductors of the second class in detail, 
when discussing the various furnace types. It may be well 
here to speak of the well-known application of a conductor of 
the second class, in the form of the filament or glower of the 


Nernst lamp, which consists of porcelain and magnesia, which 
substances are non-conductors at ordinary temperatures. The 
glower of the Nernst lamp must therefore be warmed up first, 
before it is in any way capable of taking up the lighting current. 
This pre-heating was at first accomplished with the heat of an 
ordinary match, whereas it is to-day done electrically. All 
substances which are used for fire-resisting materials for the 
lining of electric furnaces, are similar in a way to the filament 
of a Nernst lamp, i.e., all constructional material for the hearth 
or roof becomes a more or less good conductor, at the high tem- 
peratures which are prevalent in electric furnaces, and this is, 
of course, taken into consideration in their construction, as will 
be made evident later on. 

The constructional material used most, for the lining of 
electric furnaces, is dolomite or magnesite, aside from the pro- 
tecting brickwork used as a backing, and aside from the roof 
material. This material is mixed with tar and pressed into 
bricks, which are later on used in the furnace hearth, or it is 
directly tamped into the furnace. The furnace walls produced 
in this way have a small conductivity in their cold state, so that 
in such a case they may be regarded as non-conductors. They 
lose their resistance with increasing temperatures very fast, as 
is shown in Fig. 9, which shows the results of an investigation, in 
graphic form, of the resistance measurements of magnesite and 
tar rods, in relation to the temperature. The curve shows 
plainly how the specific resistance suddenly falls. This is desig- 
nated by p in the figure. It also gives a characteristic picture 
of the conductivity conditions, as they appear with conductors 
of the second class. We will recur to this matter again in due 

As we have already dealt with the prime laws of all electrical 
calculations, we will now deal briefly with the possibilities of 
different connections. This leads us to the series and parallel 
connections. Both connections will again be met in the detailed 
description of electric furnaces. As both have their advantages 
and disadvantages, it seems well that the prime difference be 
clearly stated. 


We will therefore again recur to the analogy between elec- 
tricity and water. Let us suppose we had a water-power of 
very high fall but of comparatively very little water, and that 
several water-wheels were to be driven with it. If the wheels 
could be operated at any place it would be best, with the small 

200 250 300 350 400 450 500 550 600 650 700 750 800 850 900C. 
Temperature > 

FIG. 9. 

amount of water available, that it be used time and again, or 
sausage-like in series, through the different wheels. That is to 
say, the water-wheels could be arranged at different elevations 
of the fall, and thus by dividing the pressure an equal amount 
of water could be used to drive each wheel. The small amount 
of water is similar to a small current, the high fall or great differ- 
ence in pressure is similar to a high voltage. In such a case if 
the current present can be used in the electric apparatus installed, 
but only a portion of the prevalent voltage is required, then the 
apparatus utilizing the current can be so made, that it only 
absorbs a portion of the voltage. In this case, therefore, the same 
current and same amperage flows through the various apparatus 
using it, and we consequently speak of a " series connection." 


We may, however, also conceive of a case where a water-fall 
consists of a considerable quantity of water, but has only a small 
pressure. A well-known case is on the Mississippi River at 
Keokuk, where eventually 230,000 HP, at 25 cycles, 3-phase 
current will be generated. Here it would be practically im- 
possible to use this immense volume of water in one turbine. 
We are therefore obliged to separate these large volumes of \vater 
where each part materially has the same fall, between the intake 
and the tail-race. We have here then a case where several 
turbines are arranged next to each other. This case also has 
its simile in electrical engineering, the best known case being 
perhaps the ordinary incandescent lamp, where the circuit is 
also so arranged that only a small part of the main current 
flows through each lamp, at the same voltage for each. There 
may be any number of lamps next to each other, or as we say, 
they receive their current in parallel. 

Both cases are used in electric-furnace constructions. The 
Heroult furnace is an example of a series connection, while the 
Girod furnace employs a parallel connection. 

These furnaces are described later on, but a few words here 
concerning their method of connections will not be amiss. 
Figure 10 shows the schematic wiring diagram of a Heroult 
furnace, while Fig. 1 1 shows the main features of the connection 
in a Girod furnace. It is evident, that in Fig. 10, the current 
would flow through first one and then the other conductor, 
whereas in Fig. n the two points of the circuit are connected 
together by means of two conductors connected in parallel. Of 
course Ohm's law is applicable in both cases. Accordingly with 
a resistance of r in the conductor and having a voltage e between 

its terminals, we would have a current of i = amperes. If 

the voltage e, in Fig. n, is prevalent between the points A and 
B, Ohm's law will, of course, hold for each parallel connected 

Suppose that between the points A and B, we have the 
voltage e and between these points we also have the resistances 
ri and r 2 , the resultant current being ii and i 2 , respectively. 


Then in a similar way as with the water-wheels, the main current 
will be equal to the sum of its parts, i.e., i = ii -+- i. We there- 
fore have the first of Kirchhoffs laws: "At each point of 
division the sum of all the incoming currents equals the sum of all 
the outgoing currents, or at each . point of division the sum of all 
currents equals zero." With this it is assumed that the incoming 
currents are regarded as positive and the outgoing as negative 

FIG. 10. 

FIG. ii. 

e e 

From Fig. 1 1 it follows that, i\ = and i 2 = or e = i\ r\ 

r\ r z 

and e = i 2 r 2 . From this it follows that i\ X r\ = i 2 X r 2 or we 
have the proportion ii :iz ::r z :TI, i.e., currents which flow 
parallel to each other vary inversely as the resistances of the parallel 
connected conductors . ' ' 

We will now investigate how large the combined resistance 
becomes, i.e., the resistance which is there, when both the 
parallel connected conductors r and r 2 are opposed to the current 

In order to answer this question, suppose the two parallel 
connected conductors to be replaced by a single conductor, 
having the same combined resistance r. This would not change 


the current conditions, and with the voltage e between A and B 

remaining the same, we would also have the total current i 

remaining, so that, i = -- as before. 

As demonstrated before i = ii + z' 2 . If we substitute for 
these current volumes their corresponding voltage and resist- 
ance equivalents we have: 

1 *1 = 

e e e i i i 

consequently = h or = 1 

J r r, r 2 r r l r z 

This gives the size of the combined resistance. 

= ri x r * 

In the same way this rule also holds for n parallel connected 
circuits. Designating the combined resistances again by r, we 

i i i i i 

r ri r 2 r z n' 

As the reciprocal of the resistance is designated as the conduc- 
tivity, we may define this equation by: 

"The conductivity of a combination of conductors is equal to 
the sum of the conductivities of the single conductors" 

If the n conductors are equal to each other, then 
11,11 ^ i n r\ 

r r\ r\ r\ r\ r\ n 

i.e., "The combined resistance, of n parallel connected resist- 
ances, equal to each other, is equal to the nth part of any single 

In accordance with the above, it is now possible to give an 
arithmetical example of electric-circuit conditions, as they are 
often found with electric furnaces. 


The circuit is to consist of two carbon blocks or electrodes 
in touch with a connecting iron block. The carbons and the iron 
may have equal cross-sections, and may be round of, say, 350 mm. 
diameter (about 14 inches). The length of each carbon block 
is to be 1.5 metres (about 60 inches), and the length of the iron 
block i metre (about 40 inches). 

In calculating the resistance of each part of the circuit we 

know that r = p X . In glancing at Fig. 12, we may take the 

circuit as consisting of two equal parts, which are connected in 
series in the first place, and in parallel in the second case. The 
resistance /, as shown by the figure, is composed of the resist- 
ance r c of the carbon block, and the resistance r Fe or half of the 
iron block. 


r c 

.00779 ohm. 

(Here the average value for p c is taken from 
the table on page 14, where the specific resis- 
tances of carbon are given.) 

he -5 

r Fe = PFe 


.000000572 ohm. 

FIG. 12. 

Here the average value for p Fe was also taken 
from the table. 

Even this short example shows how very 
small the iron resistance is compared with the 
carbon. For even were the carbon and 
iron, in the above example, of equal lengths, the iron resis- 
tance value would have risen to three times the value given 

( = ) , but even so the carbon would still have a resistance 
V -Soo/' 

2,500 times as great as that of iron, and this is entirely on account 
of the extraordinary differences in their specific resistances. 

Turning again to the example, we find that we know the re- 
sistance of the parts of which the circuit is composed. It is 

r' = r c + r Fe = .007790572 ohm. 


In case of a series connection the current would have to 
traverse this resistance twice, so that this resistance would be, 

r H = 2 r' = 2 X .00779 = - OI 55 8 onm - 
In the second case, i.e., with a parallel connection, we have a 
total resistance composed of a combined resistance, consisting 
of two similarly constituted parts, each having a resistance of 

r' = r c -f- r Fe = .007790572 ohm. 

As the combination resistances are alike and in parallel, we 
have n = 2 for the above case, or 

r p = = .003895 + ohm. 

It is evident, therefore, with the same conductor under 
the same conditions, but in the first case having been in series, 
and in the other case connected in parallel, that the series 
connection has four times the resistance of the parallel con- 

This extraordinarily different resistance of the two connections 
is, of course, not without its influence on the current and voltage 

conditions. This is evident from Ohm's law, where i = and 

the example before us with equal voltages gives us four times 
the current with the parallel connection, as it does with the series 
connection. Or as e = i r, and if we wanted equal currents in 
both cases, we would have to have four times the voltage in the 
case of the series connections, compared to the parallel arrange- 
ment of conductors. 

It may be well to mention here, that the above example 
only holds for arc furnaces, with series or parallel connected 
electrodes, when the electrode measurements are alike, as they 
were assumed to be in the example. 

What action is there then, when an electric current flows through 
a conductor? It has always been evident in order that the 
current may flow through a conductor that a definite voltage 
was necessary, in order to overcome the resistance. Consequent- 
ly, when an electric current flows through any conductor a 


certain work is accomplished, which must come to the surface in 
one form or another. In our case, we find work produced by 
the current, showing in the conductor again as heat. In what 
degree this heat is developed, is given us by Joule's law. This 
was established by the English physicist Joule, and experiment- 
ally determined by him. It is as follows: 

"The heat developed by a current flowing through a conductor, 
is directly proportional to the time, proportional to the resistance 
and proportional to the square of the current," or Q = C i 2 rt, 

Q = the heat generated 

t = the time the current is flowing 

i = strength of the current 

r = the resistance 
and C = a constant dependent on the units chosen. 

If the current i is measured in amperes, the potential e in 
volts and the time / in seconds, then C = .24, which has been 
determined by most accurate measurements. Therefore 

Q = .24 i"r i gram calories or, as according to Ohm's law 
e = ir, we have 
Q = .24 eit gram calories. 

If the heat is to be measured in kilogram calories, it is to be 
noted that i Kg. calorie = 1000 gram calories, so that the right 
side of our equation is to be divided by 1000. 

If the work produced is not to be expressed in kilogram 
calories, but in metre kilograms (or the equivalent of the foot lb.), 
we find that i Kg. calorie = 424.7 metre kilograms. If the 
work delivered be designated by A , we have 

.24 eit 
A = - - 424.7 = .1010 eit m. kg. 


now. 1019 = i.e., it is equal to the reciprocal of accelera- 
tion, consequently 

A = - metre kilograms. 


We usually denote the work delivered in one second as effect or 
power, and have chosen p to denote it, hence, 


p = r- metre kilograms. 

With electrical measurements we do not desire to determine 
the effect in kilogram metres, but in wits X ampers, or in short 
voltampers, otherwise known as watts, while the work in joules = 
.24 eit, or as eit is measured in watt-seconds, watt-hours or kilo- 
watt-hours, all in accordance with the measurement of time, 
be it in seconds or hours, and whether the power is to be 
inserted in watts or kilowatts. Thereby i kilowatt = 1000 

For the conversion of metric horse-power into watts, the 
following are determining factors: 

i HP metric = 75 kg. m.; as the electric power 

e i 
p = 5- in kg. = ei watt 


i.e., i m. kg. = 9.81 watt, consequently i HP = 75 m. kg. 
= 75 X 9.81 watt = 736 watts. 1 

This gives the relation between the mechanical and electrical 
units. If we arrange the determined factors in the form of 
a table, we obtain the following: 

The heat generated in a conductor by a current is, 
Q = .24 eit gram calories 
= .24 i 2 rt gram calories 
where e is to be inserted in volts 
i is in amperes 
r is in ohms 
/ is in seconds. 

The power, or the effect is 

p = ei watt or volt amperes 
= i 2 r watt. 

1 One British HP = 33,000 ft. Ib. = 746 watts. 


Here i kilowatt = 1000 watts 

i m. kg. 
i m. kg. = 9.806 watt 


i HP = 736 watts = 75 

The delivered work is given by the formula, 
A = eit watt-seconds or joules 

= i 2 rt watt-seconds or joules 
Here i watt-second = .24 gram calories = i joule 

= . 10198 m. kg. 
i watt-hour = 3600 joules = 864.5 gram calories 

= 367. i m. kg. 
i kilowatt-hour = 1000 watt-hours 

= 864.5 kilogram calories 
= 367ii4m. kg. 

i m. kg. = 2 -35 gram calories 

= 9.806 watt-seconds. 


THE effects of the electric current which interest us most, 
are its heating effects. Let us therefore consider what possi- 
bilities electrical engineering offers for the production of heat. 


In the foregoing chapter we have seen that when an electric 
current flows through a conductor, heat is developed and the 
quantity produced is 

Q .24 eit gram calories 

= .24 ? rt gram calories. 

If we for example force an electric current through an iron 
conductor the temperature of this conductor will increase. The 
heating will therefore be greater and more rapid with increasing 
current and increasing resistance of the iron conductor. Possible 
methods for making this resistance larger become evident from 
a glance at the formula recently mentioned which reads: 


r-- P - 

An increase of resistance occurs if the cross-section is reduced, 
or if the length of the iron conductor or the liquid metal bath is 

The method of heating thus explained may be called direct 
heating by resistance for the heating is affected solely by the 
inherent resistance of the metal to be heated. 

The Taussig furnace, mentioned in Chapter I, is an example 
of an electric furnace built on this principle. 

When we consider that the direct heat by resistance is not 
confined to the metal it is desired to heat and that any such 
method of heating must also naturally cause heat to appear in 
every wire used to conduct the electric current, we encounter 
the first difficulty which operates against a practical utilization 




of the direct heat by resistance for the purpose of heating metal 
baths. An iron bath is a comparatively favorable condition at 

that, for the resistance being r = p the specific resistance p is 

of considerable importance in creating the higher degree of heat. 
This point is, of course, carefully considered in electrical engineer- 
ing and it is for this reason primarily that copper is used for 
electrical circuits. Copper by virtue of its small specific resistance 
(p .018 to .019) is one of the best conductors for the transmis- 
sion of heavy currents, for it permits the use of relatively small 
cross-sections without attaining too 
high a temperature. In contrast 
thereto the specific resistance of iron, 
for example (p = .1 to .12), would be 
the deciding factor from the stand- 
point of heat loss, and with equal 
cross-sections for copper and iron we 
would have a greater heat in the 
latter in the proportion of .1 to 

With equal cross-sections for iron 
and copper, and equal lengths of the 
conductors, iron will attain a consid- 
erably higher temperature. Espe- 
cially as its heat will be further 
increased by the fact that the metal 
in the heat protective covering of the 
electric furnace will rise to higher 
temperatures by reason of its posi- 
tive heat coefficient, which in turn 
adds to its resistance. Yet the fact 

FIG. 13. 

FIG. 14. 

FIG. 15. 

still remains that the specific resistance of iron must be placed as 
being very low. With a system of pure resistance heating and 
with a resistor of iron, exceedingly high current strengths would 
be required to produce the temperatures needed in electric fur- 
naces, and the currents would have to be exceedingly strong even 
if the cross-section of its iron resistor should be made very great 


and the furnace bath cross-section very small in order to 
increase the resistance as much as possible. 

An example will serve to illustrate more clearly the condi- 
tions of this kind of electric furnace employing direct heat 

Suppose we assume that by means of pure resistance heating, 
a fluid iron bath of one ton is to be supplied with 200 kw. In 
this case the heat would be generated only by the current over- 
coming the resistance of the bath. This is the amount of energy 
with which an electric furnace of i-ton capacity would normally 
be operated. In order to make the iron bath of as large a 
resistance as possible, the molten iron is to be contained in a 
long channel of the smallest practicable cross-section. A furnace 
of this kind having a long narrow channel has also been patented 
by GIN. Figs. 13, 14 and 15 show views of this furnace, re- 
spectively longitudinal cross-section, vertical cross-section and 
a plan view. We might use this arrangement as an example. 
The size of cross-section has been taken as 10 cm. (4 inches) 
high and 5 cm. (2 inches) wide. This has been done in order 
not to reduce the channel cross-section to such a degree, that 
the furnace would be rendered inoperative. Or if the cross- 
section were made much smaller, the cooling surface would 
become extraordinarily large, and thus cause very large losses. 
This would be entirely independent of the metallurgical difficul- 
ties which would ensue if any slagging work were attempted in 
the narrow channels. The size channel chosen therefore would 
still be practically workable. 

The following assumptions are therefore made for this 
example : 

Capacity of furnace I ton = (i ,000 Kg.') 

Energy consumption A = 200 kw. 

Cross-section of bath q = 5 S Q- m.m. 

Specific gravity of molten iron, about 7.0 

Specific resistance p of molten iron, about 1 .66 

From this it follows that: 

Volume of iron = - - = 142.8 cu. decimetres. 



Length of the iron column L = 
metres = about 94 ft. 

Hence the resistance would be: 


285.6 d.c.m. = 28.5 

r = p X = 1.66 X - = .00946 ohm. 
q 5000 

As the Joule effect A = r r, we have 


= 4598 amperes. 

. 00946 

It is evident therefore that a very considerable current would 
have to be supplied. This also means large copper cables for 
bringing the current to the furnace, as those of inadequate cross- 
section would heat up too much. 

The following table shows the currents usually permitted 
in wires and cables: 


Circular Mils, (d 2 ) 
i Mil. = .oo i Inch 

Square Mils. 
(d 2 x .7854) 

Rubber Insulations 

Other Insulations 























. 65. 


66373 - 














150. ' 


















500000 . 

3920OO . 



60000O . 




70000O . 




80000O . 










The above table is for insulated wires, whereas bare conduct 
ors may carry higher current densities. 

If in the previous example for the i-ton furnace we allow 
a current density of 1.5 amperes per square millimetre (about 
1000 amperes per sq. inch), the secondary conductors leading, 
directly to the furnace would have a cross-section of 

= 3065 sq. mm. (about 4.75 sq. inches). 

This would entail .3065 X 10 X 8.9 = 27.28 kg. of copper per 
meter length (or about 19.5 Ibs. per foot length). 

Even this value shows that the leads for furnaces of the 
simple resistance type become extraordinarily expensive, and this 
is especially so for furnaces of larger capacities. 

The voltage necessary to force the required current through 
the iron bath of the i-ton furnace is 

e = ir = 4598 X .00946 = 43.5 volts. 

These give us the entire range of electrical conditions, but 
these must be considered primarily for direct current. When 
operating electric furnaces for alternating current, there would 
be certain deviations, which will not be taken into consideration, 
as direct current gives simpler equations, so that the above 
calculations are quite sufficient for the case before us. 

Considering the above circumstances, we may now establish 
the following: 

Characteristic marks of electric furnaces having direct resistance 

a. Concerning the Electric Characteristics 

As the heating occurs by means of the current passing 
through, and overcoming the resistance of the iron bath, it is 
entirely uniform at all places. 

Furthermore as the heat generated is proportional to the 
square of the current, the smallest changes in the temperature 
may be brought about by altering the voltage of the furnace. 
Such changes would be absolutely uniform throughout the 
entire bath. In this way by choosing high enough voltages, the 


highest temperature may be reached. Besides this the following 
may be noted : 

As the iron has a comparatively low specific resistance only 
at the high temperatures prevalent in liquid iron baths, it follows 
that direct resistance heating may only be accomplished by 
applying very heavy currents. For the same reason the voltages 
required for the operation of these furnaces are comparatively 

The high currents have the disadvantage of demanding 
expensive cable installations, whereas the low voltage has the 
advantage of simpler and easier insulation, and the advantage 
of eliminating all danger which might befall the furnace attend- 

The high currents can only be decreased by correspondingly 
increasing the voltage and contracting the bath cross-section, 
or by increasing the length of the bath. 

This brings us to : 

b. The Metallurgical Characteristics 

Primarily the good points here are the uniform heating, and 
the easy regulation within narrow limits at any desired high 

The disadvantages are : 

To obtain good electrical conditions it is necessary to use 
long channels having small cross-section which means large 
cooling surfaces. This is equivalent to high thermal losses 
which must be covered by expensive electrical energy con- 
sequently making the power-consumption figures very high. 

It seems that a regular operation of metallurgical process is 
precluded, as the working with slag and moreover the changing 
of slag would almost offer practically unsurmountable obstacles 
in the line channels. A uniform composition of the furnace 
contents would hence be unattainable. A lasting 'durability of 
the furnace refractories also seems practically unobtainable, 
considering that the refractory walls between the channels are 
attacked from both sides by molten iron. 

This direct resistance furnace, with its channels running to 


and fro, still fails in spite of several electrical advantages. Prac- 
tical operation has also shown this. For the one furnace of 
GIN built as here described, was an utter failure. Even so, 
we see patent applications today of similar ideas, which are 
to be discarded, as they are bound to be unsuccessful owing to 
the same inherent weaknesses. 

These direct resistance furnaces, as just described can there- 
fore not be considered for practical operation in the iron and 
steel industry. 


If we are able to circumvent the difficulties of leading very 
heavy currents to the iron bath, we may decrease the resistance 
of the bath at will if we can only increase the current strength 
to correspond. This would then allow us to use furnaces with 
large and wide hearths such as the metallurgist must necessarily 

The solution of the problem is found in the furnace type 
known as induction furnaces, which may also be called furnaces 
with resistance heating. These have the good points of resist- 
ance heating with the current being caused by induction, without 
bringing with them the disadvantage, just mentioned above, of 
the current having to be led to the furnace with immense con- 
ductors. This is what has enabled these furnaces to attain 
their great practical importance. 

On this account induction furnaces are discussed at length 
in the tenth and following chapters. It suffices to say here, 
that the induction furnace belongs to that group having a type 
of direct resistance heating. 


We shall designate all furnaces as resistance furnaces with 
indirect heating in which the iron itself is not the important 
resisting element but rather some other conductor of very low 

This conductor, that is the actual heat resisting element, is 
placed into the circuit and heated so that it can give up the heat 
generated in it, to the material to be heated. 


As we can now choose for the heat resisting body, a material 
having a very high specific resistance, the extremely heavy 
currents are no longer required, which were necessary for the 
direct resistance heating of molten iron. Consequently cable 
installation will be less expensive. 

Such furnaces with indirect heating are often used in labora- 
tories. We have for instance the type suggested by BORCH- 
ERS. (See Fig. 16.) This has carbon blocks or rods of large 
cross-section which serve as terminals between which a carbon 
rod of very small cross-section is clamped, which serves as the 
heating resistance or resistor. The material to be heated is 
heaped about the small carbon 
rod. Thus while the current is 
flowing, the carbon rod heats the 
desired material indirectly. 

In practice we find furnaces 
with similar indirect heating, but 
they are used mainly for the 
manufacture of Carborundum. 
Here we find that a tamped- 
in mass of powdered coke acts 
as the heat resisting material. 

Such designs are not used 
in the practical manufacture 
of iron, where the charge to 
be heated comes in direct con- 
tact with the heating element, 
where the latter is usually of 

FIG. 16. 

FIG. 17. 

carbon. For it is well known that iron absorbs carbon readily 
until it is finally saturated with it. It is consequently impossi- 
ble to use carbon as the heating element, not only because 
the carbon brings impurities to the iron, but primarily because 
the carbon resistance would be worn away in the shortest time 
by the iron. In place of carbon we could suggest the 
utilization of a conductor of the second class as the heating 
element. This could be of the same material as the fur- 
nace lining which surrounds the metal bath; such as dolomite 


or magnesite held together with anhydrous tar as binding 

If we were to place such a heating element or resistor in an 
iron bath we would have two parallel circuits for the current. 
In this case a very poor conductor (the heating element) of very 
high resistance would be in parallel with a very good conductor 
of very low resistance (the iron). As the currents in parallel 
circuits are inversely proportional to the resistances practically 
all of the current would flow through the iron while the con- 
ductor of the second class would almost be without current. 

Hence it is established that it is impracticable for any iron 
process for furnaces having indirect resistance heating, to have the 
heating element in parallel with the iron to be heated. 

Another possibility of indirect heating could be obtained by 
utilizing the walls of a vessel, such as a crucible, by heating it 
with an electric current either directly or indirectly. One of 
the best known of these furnace types is the HERA US laboratory 
furnace, where the heating chamber is composed of a cylindrical 
tube, into which small crucibles may be placed. The tube of 
refractory material is wound with a spiral of platinum wire or 
ribbon, which is placed in the electrical circuit, and thus its heat 
is transmitted to the furnace chamber. 

Similar methods, however, have been proposed for several 
iron processes, one of these being by GIROD. Accordingly 
several crucibles were placed in retorts composed of fire-brick, 
the bottoms of which were composed" of suitable resistances, as 
shown by Fig. 17. In order not to imperil the retort walls by 
the heat, various resistances were used for the bottom material. 
The resistance material itself consisted of carbon and silica. 
With a furnace of this kind utilizing indirect heating, a tempera- 
ture of 1400 to 1700 C. was reached. When the cross-section 
of the heating element was reduced, as shown in the sketch, 
temperatures as high as 2000 C. were attained. 

In these furnaces, which GIROD used principally for making 
ferro-alloys, he also melted steel. This necessitated 1440 Kw.- 
hrs. per ton melted. 

In the above we have an electric furnace which differs only 



from the ordinary crucible by the electrical heating. Even if 
these have several advantages, the GIROD crucible furnace still 
has the disadvantages of the small size of the common crucible, 
the difficulty of obtaining a complete uniformity from a greater 
number of crucibles, the high cost of the crucibles, and. compared 
to other furnaces, a very high power consumption. 

All these are reasons why these furnaces have not found a 
place in the iron industries. This furnace construction had to 
be mentioned here, in order to give as complete a picture as 
possible of the various electrical heating possibilities. 

FIG. 18. 

FIG. i8a. 

We still have to mention another indirect heating method, 
where the walls of the heating chamber are the heating elements 
themselves, and consequently carry the current. One of these 
designs is the Helberger furnace. This consists, as Fig. 18 
shows, of a crucible, which is placed in circuit by means of 
carbon contacts, so that the current passes vertically 


through the crucible walls. Helberger uses the ordinary carbon 
or graphite crucibles. 

Before using these crucibles they are prepared by a patented 
process which permits the current passage through the crucible 
walls only. These furnaces were originally built only for the 
handling of precious metals. As communicated by the firm of 
Hugo Helberger, the current conduction from the crucible wall to 
the metal contents was made more difficult by removing the 
graphite from the inner surface of the crucible. This graphite 
removal is accomplished by blowing air into the red-hot crucible. 
As long as the material is not molten there is no passage of the 
current. As soon as the contents becomes fluid, it gets into 
intimate contact with the red-hot lining, which acts as a Nernst 
filament would, so that some electrical energy also goes through 
the lining and directly through the bath. This action has no 
deleterious influence on the charge, as the metal for which these 
small furnaces are applicable is tapped as soon as it is molten, 
for a refining of the charge is not necessary or desired. 

If these furnaces are to be used in steel works for small 
trial melts, for which they seem excellent, carbon crucibles 
are used which are nearly always lined with a metal oxide from 
.4 to 1.2 inches (10 to 30 mm.) thick. These carbon crucibles, 
so the inventor advises, need only half the voltage of the graphite 
crucibles, a result of this being that the deviation from the normal 
working current is not so great, so that in this way it is possible 
to practically lead the current entirely through the walls of the 

The practical design of this furnace is therefore to be regarded 
as having been well done. The crucible is built together with 
a regulating transformer, as shown by Fig. i8a. The upper 
carbon contact covers the heating chamber at the rim only, so 
that the process going on in the crucible may easily be observed. 
The crucible is protected against radiation by a fire-brick cylinder. 
The clamping arrangements holding the carbon contacts are 
water-cooled. The size of the Helberger furnace is limited on 
account of the difficulty encountered when manufacturing larger 
crucibles. Yet, the manufacturers, The Helberger Co., of 


Munich, Germany, state that furnaces, up to a capacity of 
300 Kg. (660 Ibs.), are being successfully built to-day. 

When counting the various possible ways of heating we must 
not forget to mention the electric arc for this has found the 
widest application in the iron industry. We will spend much 
time in the following chapters, therefore, with arc heating and 
arc furnaces. They are mentioned here only for the sake of 


Besides the purely thermal action of the electric current the 
mill man will also be interested in the chemical action which 
takes place when an electric current is passed through a liquid. 
The best known example of this is the disassociation of water 
into its constituents, oxygen and hydrogen. This may be 
observed by passing a direct or continuous current through 
water. In so doing the well-known reaction takes place as 
oxygen is given off at the positive pole and hydrogen at the 
negative pole. 

To this belongs also the best known electro metallurgical 
application of electrolytic action for the smelting of aluminum. 
According to the method first proposed by Heroult, the clay is 
melted by the action of arc heating, and simultaneously the 
molten mass is separated electrolytically in such a way that the 
aluminum is freed and collected at the negative pole, whereas 
at the hanging positive carbon electrode the oxygen is set free, 
and, together with the carbon of the electrode, escapes as carbonic 
acid gas. 

These examples are sufficient to show how chemical action 
may be brought about by the electric current. In this instance 
it is to be observed that this action only occurs when direct 
current flows through the electrolyte to be separated. If on the 
other hand alternating current is used, where the current direction 
is constantly changing, then no electrolytic action can take place. 
For supposing we had an apparatus for the dissociation of water, 
which was supplied w r ith alternating instead of continuous current. 


Then during one moment with the current in one direction, we 
would obtain oxygen at the electrode, and during the next 
moment with the reversed current direction we would receive 
hydrogen. It is evident, therefore, that electrolytic effects do 
not arise when alternating current is used. 

From this it also follows that molten iron masses of electric 
furnaces, through which current passes, are not subject to any 
chemical action as long as alternating current is used. We 
could in any event, as in the above example, assume a momentary 
chemical action, which however would be reversed in the next 
moment, for even though it appears momentarily, it does not 
come into play as far as the metallurgical process is concerned. 
This assumes that the reversal of the chemical effect is not 

When using direct current in iron baths, electrolytic action 
naturally occurs, by which iron sulphides and iron phosphides 
may be separated. These suggestions have also found their 
way into the patent office. 

Electrolytic actions may however be positively harmful for 
carrying out metallurgical processes. According to CONRAD 
(see Stahl u. Eisen, 1909, p. 796) we obtain a purer product when 
using alternating current for the manufacture of ferro silicon, 
than when using direct current. For when using continuous 
current the impurities of the charge are reduced, such as calcium, 
aluminum and other metals, which then find their way into the 
final product. 

The designers of the electric furnaces for the iron industry 
today use alternating current exclusively, because they fear the 
undesirable influences in the charge due to direct current. It 
may not be out of place to quote the words here of an ardent 
supporter of electric furnaces for the iron and steel trades. We 
quqte, therefore, from Prof. Borchers and his address in 1905, a 
translation of which might be called: 

" Electrolytic effects were not sought in most reduction and 
melting tests, whether they endeavored to produce pig iron, or 
make steel, even though these electrolytic effects are nowhere 
entirely eliminated. This was particularly so in the arc proc- 


esses of earlier periods, where testers positively failed, when 
endeavoring to produce irons low in carbon. If for instance the 
iron to be smelted makes one pole of the arc, and carbon blocks 
the other, then the iron absorbs carbon equally well, whether 
direct or alternating current be used. We know that an arc 
between two carbon poles carries carbon vapor across from one 
pole to the other. For the evaporating point of carbon deter- 
mines the arc temperature. If one of the poles consists of iron, 
then the only material remaining for the other pole is carbon, 
provided direct arc heating is used. In this way the iron will 
gradually become saturated with carbon even with alternat- 
ing current. Though we assume that the carbon in the arc 
wanders only from the anode (positive pole) to the cathode 
(negative pole), it is evident that the carbon separates itself in 
its solution in such a way, during a current wave, in going from 
the carbon pole to the iron, that only a small part of it would 
return during the current alternation." 

We perceive, therefore, that we must guard against the 
harmful absorption of carbon by suitable means, when using 
furnaces operating with carbon electrodes, even when working 
with alternating current. 

This is accomplished today by interposing a layer of slag 
between the arc and the iron, in all furnaces where the arc im- 
pinges directly on the metal. This slag is then, to be sure, acted 
upon in a reducing manner by the arc, yet the iron is protected 
from any union with carbon. 

In accordance with the foregoing, then, we may definitely 
establish that no electrolytic action takes place in the great 
majority of furnaces, which operate exclusively with alternating 



In the construction and operation of electric furnaces we 
have to take into account the motor effect of the electric current 
as well as the thermal and chemical effect. 

It is just as easy to transform motion into electricity as the 
reverse, as is very evident from the wide application of the 
electric motor. 


It exceeds the limits of this book to explain the motor phe- 
nomenon in detail, still in order to understand and have a correct 
opinion of the possible and impossible motions of the molten 
metal in electric furnaces, it appears desirable at least to discuss 
briefly the reasons for the motion phenomena. 

It is well known that an ordinary magnet attracts a piece of 
iron brought into its vicinity, and that motion is caused by means 
of this magnetism. It is equally well recognized that two magnets, 
like magnetic poles, repel one another; while unlike poles attract 
one another. 

We also speak of lines of force, which surround the space 
near a magnet, and it is to these lines of force issuing from a 

magnet that we attribute the 

/ / ^- -^ x\ \ distant magnetic effect. Sup- 

/ / \\ i pose we have two magnets as 

Fig. 19 shows with their like 

:V=j|s Nf-r poles laid next to each other. 

7?\' '/rr If we draw the path of the 

^ f rce as shown we 

r^js ' Np may define the repelling action 

y/;^\ x x y / '']f\\\\ of like poles, by saying: 

'' > \ x \ x ^r~ ""'^/' I } ^ Lines of force having the 

\ \ x ^-^ _______ ~*^'/ / same direction repel each other; 

those of opposite direction 
attract each other. 

We also obtain motion phenomena, therefore, in accordance 
with this rule, as a result of two magnets acting on each other. 
But we will also have these motion phenomena, if a stationary 
current carrying conductor is brought near an ordinary suspended 
magnet. In order to do this we may set up an easily movable 
magnetic needle in its case, and directly above it stretch a wire, 
which may be connected to a source of electricity. As soon as 
the current is switched on, the needle will endeavor to set itself 
at right angles to the wire. The size of the deflection is a direct 
measure of the current passing through the wire. We find that 
the deflecting power decreases as the conductor is moved away 
from the magnet parallel to itself; that the direction of the 



needle is reversed when the wire is under instead of above the 
needle; and that at every position the deflecting power is pro- 
portional to the current. 

These phenomena prove that a current carrying wire is 
surrounded by lines of force throughout its whole length, whose 
density is greater in the immediate vicinity of the wire, which 
decreases as the distance from it (the wire) increases. Accord- 
ingly we may imagine the fields of force of a current carrying 
wire about as shown by Fig. 20. It is assumed here that the 
conductor pierces a sheet of paper. On this are drawn the lines 
of force as they would appear when the current flows. The proof 
of these lines of force existing 
concentric to the conductor may 
easily be had, if a glass plate is 
used in place of the paper, which 
is strewn with iron filings. If an 
electric current is then sent 
through the wire, which pierces 
the plate, the iron filings will 
arrange themselves in direction 
and density, in accordance with 
the lines of force. The direction 
of the lines of force may then be 
established in compliance with 
a single rule: // the current 
carrying conductor is grasped in 
the right hand so that the out- 
stretched thumb indicates the direction of the current, then the lines 
of force will encircle the wire so that they would issue from the ends 
of the remaining fingers. 

If we now return to the first test, in which a movable magnet 
was brought into the magnetic field of an electric conductor, 
then in diverting the magnet we have a motion phenomena, 
pursuant to mechanical power, which appears between electric 
currents and magnets. 

We can now go a step further and replace the second magnet 
by a conductor, through which current flows. Even then certain 


FIG. 20. 


phenomena will be observable, provided one of the conductors 
carrying current is movable. For until now, we have found that 
the motion phenomena are the result of magnetic fields which 
mutually affect each other. As we have also seen that each 
conductor carrying current has its own magnetic field, then, in 
accordance with the foregoing, the appearance of mechanical 

power is unavoidable, 
between current carrying 
conductors lying closely 

Accordingly we may 
immediately determine the 
direction of the motion. 
Suppose we have two con- 
ductors both of which 

carry current going in the same direction, as shown by Fig. 21 
at a and b, here the current would be flowing toward the reader. 
According to the foregoing rule the direction of the lines of force 
is quickly determined and is shown by the arrows. We see, 

FIG. 21. 

FIG. 22. 

therefore, in the space between the two conductors that the 
direction of the lines of force are opposite to each other. As 
the lines of force of opposite direction attract each other, we may 
say relative to the current: "Currents of like direction attract each 
other " and "currents of opposite direction repel each other" 



From this it follows that crossed currents and their con- 
ductors (as shown by Fig. 22) endeavor to arrange themselves 
parallel to each other, and in such a way that the current 
in both flows in the same direction. That is, the movable 
conductor a tries to assume the same direction as the stationary 
conductor b. 

The case is also very interesting where one current flows 
vertically to the other, as shown by Fig. 23. Here the circles or 
dots represent the points of the arrows, which indicate the 
direction of the lines of force, while the crosses represent the 
ends of these arrows. 

As lines of force in the same direction repel each other, and 
those of opposite direction attract, then the movable conductor 
a will endeavor to move in the direction as shown by the arrow. 


I . 






o o o o o o \ 

' o o o o o o 

+ + + +++ + + -* 4- -4- + 

FIG. 23. 

In place of the above simple case we may throw some light 
on the possible motion phenomena in arc furnaces, which may 
arise due to the electrical conditions which have their electrodes 
pointed directly against the bath. 

We may have the effect of two or more currents acting 
on each other. In this case one of the conductors, namely the 
molten metal, may be regarded as being movable, within certain 
limits. For the molten conductor may be mechanically, com- 
paratively easily influenced, even if only within the limits of the 

The conditions are also very similar with induction furnaces, 
excepting that in the place of the one solid conductor, we have a. 


coil of many turns. Figs. 24 and 25 show how the lines of force 
act: in the former case with the turns wound far apart, and 
latterly with the turns wound closely together. It follows, 
therefore, that coils such as these are surrounded by like lines 
of force as common rod magnets would be, and thereby the laws 
are known which govern the motion phenomena between active 
coils and single conductors. 

Aside from the above explanations it still seems desirable 
to mention a very special motion in molten conductors, through 

which current is passing, but 
which only arises in certain 
cases. This is the so-called 
" pinch effect." 

() (0) ()()() According to an address by 

~ -~ CARL BERING before the 

Canadian meeting of the 
" American Electrochemical So- 
ciety," in May, 1909, this pinch 
FlG - 2 4- effect occurs when a continuous 

or alternating current flows 

through a molten conductor. Then this conductor endeavors 
to contract itself in the line of its cross-section under the 
action of electro-magnetic forces. The contracting force is 
only small, when the current density is low, but grows with in- 
creasing current density (amperes per square millimetre or square 
inch), and can, in' extreme cases, become so large, that the cross- 
section at the contracting point may decrease to zero, thereby 
interrupting the current. The contraction primarily occurs at 
such places in the hearth which have 
already been contracted owing to occa- 
sional irregularities when tamping the 

lining material in place. The fluid FIG. 25. 

column of metal conducting current is 

therefore interrupted at the weakest place in its cross-section, 
exactly as a rope breaks at its weakest place. In addition to this, 
there is a depression where the cross-section diminishes, and on 
this slanting fluid conductor, particles of slag and the like are apt 


to follow, which then cause a further increase in the current 
density, as the conductivity of these impurities is less than that of 
the molten metal. Furthermore this would be caused at places 
where the cross-section is already weakened, so that thereby the 
actions of the pinch effect would be still further increased. 

This course of things, as pictured above, does not take place 
in any deleterious or unpleasant fashion with electric furnaces 
as they are used for the most part in the iron industry today. 
As the pinch effect only appears with comparatively high current 
densities, we find that it does not occur at all in arc furnaces. 
But it causes various motion phenomena with induction furnaces, 
as we shall presently explain. Motion phenomena which are 
entirely desirable and advantageous for the working of metal- 
lurgical processes, may be brought about by artificially narrowing 
the cross-section of the bath to accomplish the required result. 
These, by their very nature, 
would in no way endanger the 
electrical furnace operation. 

An explanation of the appear- 
ance of the pinch effect may be 
had, if we assume that the FIG 26 

fluid mass is composed of many 

parallel connected conductors, which are all leading like direc- 
tional currents through them. As currents having the same 
direction attract each other, the foregoing sentence is applicable 
here, for in a measure it defines the contracting effect. 

In electrical furnaces, where the pinch effect causes motion, 
the liquid seems to be driven along a straight line from the 
middle and the centre axis toward the ends, so that the fluid 
mass is lower in the centre than at either end. The weight of the 
molten metal then causes a flow from the higher lying parts, 
toward the lower middle section, as shown by Fig. 26. Here 
the molten conductor is considered to be cut vertically, in the 
line of the horizontally running current. Without going into 
further details, it is evident that the motion due to the pinch 
effect causes an intensive mixing of the charge. This occurs as 
Jong as the correct agitation is maintained within the desired 


limits, for the motion can only advantageously effect a rapid 
chemical reaction, which is needed between the iron bath and the 
slag. Besides this the quality of the steel can only be bettered 
by the greatest possible uniformity which is brought about by 
this circulation. 



IN the previous chapter it was shown that direct current, 
due to its chemical action, is totally unadapted for electric fur- 
naces as used in the steel industry, and alternating current is 
therefore used exclusively to operate electro-steel furnaces. 

The difference between direct current and alternating current 
is that in the former the current is always flowing in the same 


FIG. 27. 

FIG. 2/a. 

direction, whereas in the latter it changes its direction con- 

The required time for one directional change is called the 
period and is designated by T. Fig. 270 shows a complete wave 
or cycle. In this figure one cycle, therefore, takes the time, T, 
which is necessary for the current to swing through a complete 
wave. Hence, one complete cycle goes from through to the 
positive maximum, and from zero to the negative maximum 
and back to the zero point. 

For all practical purposes we can assume that the usual 
alternating current generator gives a sine wave for its electro- 
motive force. This being the case, it only remains to show how 



a sine curve is constructed, and to draw another diagram next 
to Fig. 2ja showing these relations in alternating current 

If (as in Fig. 27) we let the radius or radius vector equal the 
maximum voltage reached in this sine curve, and designate this 
maximum value by e, the various instantaneous values of the 
sine curve by e f , then: 

e' = e sin a 

i.e., for every angle a, the ordinates of the sine curve give the 
corresponding instantaneous potential values as indicated by 
Fig. 27. In the above equation in place of the angle, however, 
we can substitute for it the value of the angular velocity and 
obtain : 

a = m t 

(similar to the equation, distance = speed X time), and as the 

d a . 
angular velocity w = -y- hence 

e' = e sin m t 

where / is the time taken by the radius vector until it has passed 
through the angle a after leaving the zero or starting point. The 
whole time corresponding to one cycle is T and the corresponding 
angle is 2 TT, and by substituting these values in the previous 
formula, since (a = 2 IT and / = T) 

2 TT = m T from which it follows that 

2 7T 

m = ~' 

It is customary to speak of cycles per second or frequency, 
and as the time of one cycle is equal to T, the frequency v is 


If we substitute this value in the formula containing m, we 
get m = 2 TT v. 

We speak of an alternating current of, say 25 cycles, when 
this current makes 25 waves each second. 

The more or less frequently varying direction and strength 
of the current depending upon the cycles per second, or frequency, 


has a particular bearing on the functions of alternating current 
circuits. In order to understand these, the so-called induction 
will next be briefly described. This seems necessary because a 
clear conception of the induction phenomena is important, in 
order to understand the induction furnaces which will later be 
discussed in detail. 

We obtain an inductive action, for instance, when an electric 
conductor is moved through a magnetic field so that magnetic 
lines of force are cut. If we connect the ends of this conductor 
with a measuring instrument we obtain a deflection, showing 
the presence of an electric current produced by induction. This 
current is called induced. 

We therefore say, " // a conductor is moved in a magnetic field 
so as to cut magnetic lines of force, an electro-motive force is produced, 
which will cause a current to flow provided that the conductor has 
its ends closed so as to form an electric circuit. The electro-motive 
force and also the current become larger, as more magnetic lines of 
force are cut in a given time." 

It is evident that it makes no difference in which way the 
magnetic field is produced, because it is only necessary for the 
conductor to cut lines of force. It is, therefore, immaterial 
whether the conductor is moved through the field of a permanent 
magnet or through the field of an electro-magnet. It is even 
sufficient to move it near a wire through which a current is 
flowing, because this wire is surrounded by lines of force. 

Until now we have assumed that we have moved the con- 
ductor in which a current is induced. Instead of that we can 
move the magnet and hold the conductor; still, as in that case, 
an electro-motive force is also generated, due to lines of force 
being cut. We may finally place two conductors side by side, 
and if we pass a current through one of them, it will generate a 
magnetic field, the lines of force of which will cut the second 
conductor. If the current is interrupted, the lines of force dis- 
appear, only to reappear instantly upon the current being again 
made. We therefore have a field of constantly changing lines 
of force and a conductor located in this field. Hence an e.m.f. 
is induced in the second conductor, exactly as when a magnet 


approaches a conductor from an infinite distance and then recedes 
again to an infinite distance. 

The alternating current changes its strength continually, and, 
as we have seen, it increases twice during each period or cycle 
from zero to a maximum and consequently decreases from that 
point again to zero. As a result of this, a conductor carrying 
an alternating current is surrounded by an alternating magnetic 
field, which induces e.m.f., or currents, in all conductors within 
its field. 

The current-carrying conductor itself thus lies in an alternat- 
ing field and, from what has been said, it is evident that an e.m.f. 
will be induced in this conductor by its own field. This action 
is called self-induction, and the current generated thereby is 
called the self-induced current. 

This self-induced current always flows in the opposite direction 
to the current which produces it. If the primary current, for 
instance, flows to the right, the induced current will flow in the 
opposite direction, or to the left, in the same conductor. The 
self-induced current for this reason does not exist, as the effect 
is to weaken the primary current. If voltage is applied to a coil,, 
therefore, the current does not immediately reach its maximum 
value, but does so only after a certain time-interval has elapsed. 
The highest value is reached after the lines of force are no longer 
on the increase. We therefore say the current lags behind the 

It should be remembered that we obtain the instantaneous 
values of the voltage as the projections of a rotating radius 
vector. Therefore, we can likewise get the instantaneous values 
of the current as projections of a radius vector of a different 
value. We then obtain the lag of the current behind the voltage, 
and draw this lag out in the form of a definite angle. This angle 
is then the measure of the lag. Time difference between current 
and voltage we call phase displacement, and the angle which 
the radii vector of the current and voltage make with each other 
is called the phase angle. The letter < has been commonly chosen 
to designate this angle. 

We have for instance the vector diagram Fig. 28, which 


pictures the relations as they might be in an alternating current 
circuit. We have only to imagine the radii vector as rotating 
about as a centre, to obtain at any time the corresponding 
values of the current and voltage by drawing the vertical pro- 
jections of their respective radii vectors. 

Sometimes this vector diagram is drawn even more simply, 
see Fig. 29. 

On what conditions now does this phase displacement depend? 

We have already seen that this phase displacement is a 
result of the self-induction. Therefore, the greater the current 

FIG. 28. FIG. 29. 

and the faster the changes, the greater is the change in the 
corresponding magnetic fields, in a given time. The frequency 
therefore, has a large influence on the phase displacement. 

Besides this, the self-induction is also dependent on the type 
of conductor used and its position relative to other conductors. 
The factor which designates these conditions is called the " co- 
efficient of self -induction." The mathematical symbol for this 
is L" 


We therefore say: 

The electro-motive force of self-induction is proportional to the 
coefficient of self-induction, and to the rate of change of current per 
second, or the frequency. 

If we compare the relations in an alternating current circuit 
with those in a direct current circuit, we see, in the latter case, 
that it takes a definite voltage to force a current i, through the 
resistance r, and, according to Ohm's Law, we have 

e = i X r 

If it is desired to send an equal alternating current, i, through 
a coil, it also takes a certain voltage, 

e r = iX r, 
to overcome the resistance. 

We have to take into account, though, that with alternating 
current an electro-motive force due to self-induction is generated, 
which is always in the opposite direction to the impressed electro- 
motive force. 

In order, therefore, to obtain the desired current i, we need 
not only the voltage, e r i r, but also an additional pressure 
e L to overcome the electro-motive force of self-induction. Hence, 
the total voltage necessary for an alternating current is, 

e = e r + e L . 

The alternating current voltage e is composed of two different 
pressure waves. These waves are displaced by an angle of 90 
or ^ of a period, which can easily be shown by a short mathemati- 
cal demonstration. 

The above sentence in italics regarding self-induction, is 
mathematically expressed as follows: 

_ di 

e ^ L Tf 

Furthermore, we know that for a sine wave, the formula for 
an alternating current at any instant is: 

i' = I sin m t, 

exactly as the sine wave for the voltage gave 

e' = e sin m t. 


If we substitute another value for i in the equation 

e L = L 



d (I sin m t) 

we obtain 

= ml cos m t 

dt dt 

e f L = (m X I X L) cos m t. 
We have e r = ir or the instantaneous value 

e f r = (l r) sin m t. 
The total voltage is, therefore, 

e' = e f r + e' L = (l r) sin m t + (i m L) cos m t 
and as cos m t = sin (m t + 90) 
it is evident that the voltage 
necessary to overcome the coun- 
ter electro-motive force of self- 
induction is 90 ahead of the 
e.m.f. necessary to overcome the 
ohmic resistance. 

From this it follows that 
these two e.m.f.'s are not to be 
added arithmetically but geo- 
metrically. If we draw this as 
shown in Fig. 30, we have : 

A = maximum value 

of the current = i 
B = e r = I X r 
O C = e L = Im L 

The resultant of the two e.m.f.'s is graphically shown as O D, 
and from it we obtain the total voltage 

e = e r + e L . 
From the figure then we have : 

e = v 7 * 2 + e r 2 

FiG. 30. 

+ I 2 

= i V r + m~ L 2 

It also follows that tan = 

m L 

when <f> is the phase angle 


between the current and the voltage. 

It is worthy of mention that from the equation 

e = i \r + m" U 

it seems as though the self-induction apparently increases the 
resistance. Hence, the expression 

v 2 + m 2 L 2 

is also called the " apparent resistance" of an alternating current 
circuit. In order that there shall be no mistake regarding the 
values which are indicated by measuring instruments in alternat- 
ing currents, it is well to emphasize here, that^so far we have only 
mentioned the instantaneous and maximum values. As a matter 
of fact, neither of these values is indicated by the usual alternat- 
ing current instruments. These values have only been used to 
more clearly state the relation in a.c. and to make them easier 
to understand. The instantaneous and maximum values are 
therefore only of theoretical interest, whereas the a.c. instruments 
indicate a so-called "effective value." This is obtained from the 
previous formulas and figures by dividing the maximum values 

__ p 

by \/2. Hence, the effective value of the voltage is e = = and 


the effective value of the current is 


We can therefore regard the diagrammatic figures as representing 
the effective values, as these only differ from the maximum values 
by a constant factor. 

If we now return to the phase displacement between the 
current and voltage, we find the question becomes of the greatest 

What influence has the phase displacement on the power 
computation ? 

It was shown in Chapter II. that the power in watts is equal 
to the product of the current and voltage, that is p = e X i. 
Unless the so-called power factor, which will be later explained, is 
unity, this last equation is only applicable to direct current. 


Whereas for alternating current the formula becomes, 

p = e i cos 4>. 

In this equation e is the effective voltage, i the effective current 
and cos 4> the power factor. 

In alternating current circuits we call the product e X i the 
apparent power. It is measured in volt-amperes or kilo-volt- 
amperes = 1000 volt-amperes. The product e i cos < designates 
the real or effective power and is measured in watts or kilowatts. 
To verify the equation for the true power really goes beyond the 
limits of this book. For those, therefore, who are interested in 
this paragraph, it is added in an abbreviated manner. 

The equation for the instantaneous energy is 

p' = e' X i'. 

T . 
The work done in ^ a period during the time is then 


A= 2 e'i'dt 


and from this we obtain the mean value of the energy. 

JL L .! 

p= T_T* e 'i'dt = 2 e'i'dt 


by substituting the values i f = I sin m t and 

e f = e sin (m t -f- $) 
and by completing the integration, we obtain, 

p = - cos and as 

e I 

= = e and -7= = * we g et P = e * cos < 

V2 V 2 

From this it follows that, providing the voltage and power 
remain unchanged, the current decreases with an increasing 
power factor. As the current strength determines the cross- 
section of the electrical conductor, it naturally interests us to keep 
the current down, i.e., we strive to obtain the highest possible 
power factor. 

From the above power equation, it follows that, when 
cos< = i, 


p = e i and the angle </> = o. The other limit is when 
cos = o or the angle = 90, then the power, p = o. A low 
power factor, therefore, corresponds with a large phase dis- 
placement. The meaning of the above may best be enlarged 
upon by an example: 

Suppose the electrical circuit contains a coefficient of self- 
induction L = .002 henry 
a resistance r = .01 2 5 ohm 

a frequency v = 50 and therefore m = 2 IT v = $14. 
voltage e = 150 volts. 


m L 314 X .002 

tan = - = - = 50.24. 

r .0125 

The angle corresponding to this value is then 88 50' or nearly 
90. Hence cos is nearly zero. 

The relations are graphically shown in Fig. 31. This shows 

im L= 

3 volts l\$ 

FIG. 31. 

that e and e L almost coincide, so that e L practically equal e. 
The current is then 

e 150 

^ = 7- = , = 240 amperes. 

mL .628 

Consequently ir = 240 X .0125 = 3 volts = ecos0 

i X e cos0 = 3 X 240 = 720 watts. 

With the same current but with cos = i , we would have ob- 
tained instead of the above, the power p = 240 X 150 X i = 
36000 watts. 

This example shows us plainly how impossible it is to judge 
the power in an alternating current circuit by merely reading 


the ammeter and the voltmeter, as these two instruments do not 
in any way indicate what the power factor is. We therefore em- 
ploy a special instrument to measure the power, a so-called 
wattmeter, which indicates the watts or kilowatts, directly, 

i kilowatt = i kw = 1000 watts. 

As the example showed that the voltage necessary to overcome 
the e.m.f. of self-induction (i.e., the vector e L ) is without any 
influence on the actual power in other words, it delivers no 
power which can be measured in watts we therefore call this 
vector the wattless component, and the vector e r = i X r is 
called the watt component of the voltage. 

Up to the present we have divided the voltage into two 
component parts. The one being the watt component e r = ir 
which coincides with the direction of the current, the second being 
the wattless component e L = i m L which is in quadrature with 
the former. 

The power p = e i cos = i ( e cos <) where e cos < = e r ; 
that is, the power is obtained by multiplying the two unidirec- 
tional vectors or forces (i and e r ). (See Fig. 30.) 

Instead of separating the voltage into two components, we 
could have also separated the current into two forces at right 
angles to each other. This separation can be done in such a way 
that one force falls in the direction of the terminal voltage, and 
being multiplied with this, it gives the resultant power, while the 
other force is perpendicular to the first one. 

The equation for the power, 

p = e i cos can be written 
as p = e (i cos 0) = e i r 

where i r is the watt component of the current and equals i cos <. 

Taking then the values of the example as chosen, we obtain 
Fig. 32. The directional precedence is given by the curved 
arrow. Here the total current, i, is shown as lagging behind the 
voltage by the angle 0, similarly to the previous example. As 
the angle is approximately 90, then i and i m almost coincide 
and the wattless component of the current is 
i = i m = 240 amperes 


whereas the watt component of the current i r approximates zero. 

It is therefore apparent that the total current i m is only 
present in order to generate the e.m.f . of self-induction. In other 
words : 

It is the wattless component of the current which generates 
the lines of force. That is why this wattless current is also called 
the magnetizing current, and this is why it is designated by i m 
in the accompanying figure. 

It follows, therefore, from all which has been said of the power 
factor that: When figuring the size of electrical conductors, 
the apparent power should always be the determining factor, i.e., 
the product e X i or current X voltage, in other words, the 
kilo-volt amperes. On the other hand, the power of the prime 

FIG. 32. 

mover only takes into account the actual power, that is the 
product e i cos </> or the actual kilowatts. In other words, a poor 
or low power factor means expensive lines and electrical machinery , 
whereas it has no influence whatever on the prime mover. 

It is apparent, therefore, that it is in the interests of an 
inexpensive installation to have an acceptable power factor. It 
is to be noted, however, that in the ordinary power houses, the 
power factor varies between .6 and .8, depending on the sizes of 
the motors used and at what load these are operating. These 
values are, therefore, a guide indicating whether or not we have 
a good power factor. 

Quite independent of the current lag, we may have induction 
phenomena which will call forth other and more disagreeable 
actions than those shown, and as it is the object in designing and 


operating electric furnaces to avoid these troubles, we will 
mention them briefly. 

We have seen that an alternating current in a conductor will 
generate another alternating current in any conductor if the sec- 
ond conductor only lies in the magnetic field of the first conductor. 

We therefore obtain currents in all conductors which lie 
in the magnetic field of another conductor, and these currents 
may cause considerable power losses under certain conditions. It 
would lead us too far if we were to occupy ourselves deeply with 
these phenomena. On that account only those possibilities will 
be mentioned which lead to these power losses in electric furnaces, 
and the remedies which help to overcome these losses. 

In the first place there are the induced currents themselves, 
which may engender considerable losses. As these induced 
currents are generated in every conductor which is parallel to 
the main current, they may cause great losses when the conductor 
carrying the induced current is short-circuited. It is therefore 
necessary to avoid all designs in which, for example, an iron beam 
would follow a main conductor, so that it would then be short- 
circuited on itself. This condition is to be considered only when 
very heavy currents are present as is altogether the case with 
electric furnaces. But even here these actions may be avoided 
by carrying the incoming and outgoing conductors close together. 
In this way the magnetic fields for instance those made by the 
two conductors of a single phase circuit are then neutralizing 
each other, so that we have no action on parallel lying and closed 
iron parts. There are, however, currents induced in every 
metal part which is near an alternating current carrying con- 
ductor. These metallic parts provide splendid conductors for 
the current through which the current may be short-circuited, 
so that under certain circumstances a metallic piece of that 
kind may reach really unlooked-for temperatures. We call 
these eddy or Foucault currents. They are particularly preva- 
lent when the metal in question is magnetic, that is, a good con- 
ductor for the magnetic lines of force. There would be consider- 
able losses, for instance, in the cooling chambers used in electrode 
furnaces, to cool the electrodes, if these were made of cast iron 


or cast steel, as both of these materials carry the magnetic flux 
better than air. We are, therefore, obliged to make these cooling 
chambers out of copper or red brass, as both of these materials 
are non-magnetic. 

Another method used to lessen these eddy current losses, 
is to greatly subdivide the metallic parts in which these eddy 
currents might appear. Transformer and dynamo armature 
cores are examples. These cores are built up of sheets as thin as 
,5 and sometimes only. 3 mm. (.02 to .012 inch). 

Finally we might also have the case where a good magnetic 
conductor, one of low magnetic reluctance, entirely surrounds an 
electric conductor. If the magnetic conductor should have a 
considerable cross-section, then certain power losses arise, due to 
the constant demagnetizing influence of the alternating current. 
This loss is known as the hysteresis loss. For this reason, there- 
fore, we also avoid surrounding electrical conductors with good 
magnetic conductors in electric furnace construction. 

It seems well to mention that besides single phase alternating 
current, polyphase (2 or 3 phase) alternating current is also used 
to operate electric furnaces. In order to understand these 
power circuits, we will add the following: 

Three phase current is visually distinguishable by having 
three lines which conduct the current from the source of supply to 
the apparatus using it. Whereas with single phase current there 
are only two lines, one line to lead the current to the destination 
and one return wire. 

As the name three phase implies, we use three conductors 
and handle three currents in this power transmission. The 
vector diagram shows us this the plainest, i.e., the relations be- 
tween these currents and what the relations are between the 
different values occurring in three phase power transmission. 

Fig. 33 shows us three vectors which are separated 120 from 
each other. These vectors indicate the direction of the current 
as they are actually generated in 3-phase machines and actually 
consumed in 3-phase apparatus. If we add these currents 
geometrically, as shown in the figure, we observe that the geomet- 


rical resultant of two current forces always equals the third 
current. This explains why only three lines are necessary to 

FIG. 33. 

FIG. 340. 

FIG. 34. 

conduct a 3-phase current, of which the third conductor may be 
looked upon as a return wire for the other two. This presupposes 
of course that the current in each direction or phase is of the same 


The coils of the generator or those of the power consuming 
apparatus which are built for 3-phase current, may be connected 

FIG. 35- 

FIG. 35a. 

FIG. 36. 

in two different ways with each other. Fig. 34 shows the so- 
called Star or Y connection, in which the ends of the coils of the 
generating or receiving apparatus are connected together at the 


neutral point A, whereas Fig. 35 shows the so-called Delta con- 
nection in which the single coils are connected in series, and the 
connecting points of the coils are led off to the power mains. 

With the Star or Y connection we may have either the volt- 
age of one phase or the resultant voltage of two of the phases. 
The first is the potential between the neutral point A, Fig. 34, 
and the end of one generator coil, as shown by the connections 
Aa = Aa2 = Aa 3 = E. The other voltage is the resultant of 
two of these coils and is across the points a 0%, #2 a^ a 3 a, and this 
resultant voltage is designated by e. If in these Star connections 
the phase voltages should be different, there would however be 
no difference between the currents flowing in the generator coils 
and on the line. If / = generator phase current, and i = line 
current, then / = i, as is evident by consulting Fig. 34. 

If we, however, view Fig. 35, we instantly perceive that the 
phase voltage and line voltage are equal to each other or E = e. 
On the other hand we have different values for the current per 
phase and the line current. With 3-phase currents for electric 
furnaces the Y connection is mostly used. What is the relation 
between these two voltages? 

If we have the phase voltage E, we may obtain the resultant 

voltage by taking the geometric difference between any 2 phase 

voltages. If we refer to Fig. 36 we see that, sin 60 = E and 

therefore the resultant voltage 

e = 2 E sin 60 
= V~$E = 1.73 E. 

In the same way it may be shown for A connection that 

* = v'J I 

These relations must be known in order to clearly understand 
the power in 3-phase circuits. We can imagine the 3-phase 
power being equal to the sum of power of the 3 single phases. 

We then obtain, 

p = EI /i cos + E 2 h cos + E z 7 3 cos 
and as we assume that the separate phases are balanced or equally 


loaded we may write that 

p = 3 E I cos <. 

We saw for the Y connection that e = \/$ E and i = I. 
If we substitute these values in the power equation, we have 

e /- 

P = 3 /~ - l cos = v 3 e l cos 

i.e., we obtain the power in a 3-phase circuit by multiplying the 
current by the voltage by the pow r er factor and the product by 


Finally, it may be said that we have wattmeters which meas- 
ure the total power, p = \/3 e i cos in 3 phase circuits. It is 
therefore an easy matter to determine the power factor in a 3- 
phase circuit, provided other instruments give the values of the 
current and voltage of the 3 balanced phases. 

We therefore have 


cos <j> = - 


In the above p = total power in either a Y or A connection circuit 
as measured by a wattmeter, i = the current in each line, 
sometimes only measured by one ammeter, and e = voltage as 
measured by the usual a.c. voltmeter. 



BEFORE we deal with the furnace designs now largely used 
for steel making, it may be well to discuss a few general questions. 
An understanding of these is of great importance in order that 
we may correctly judge an electric furnace. 

First and foremost the question arises: 

Why has the steel industry in general an interest in electric 
furnaces, and what advantages does the electric furnace offer 
compared to the existing metallurgical apparatus? 

It is obvious that the advantages will have to be of some 
moment, if the iron masters are to discard or supplement their 
hitherto satisfactory methods of procedure. 

It behooves us then to consider first the proved and peculiar 
heating effects derived entirely from electricity. We find the- 
following characteristics : 

1 . The use of electricity as a heating agent makes an extraor- 
dinary and quick heat possible, which same is impossible with 
any system of gas heating. Here it may be noted, that before the 
introduction of the electric furnace into the steel industry, it 
was only possible to make refractories stand temperatures of 
2000 C., whereas we may now reach any temperature up to 
3500 C. in the electric furnace. 

2. With the aid of electrical control the heat can be regulated 
most accurately, so that the charge can be brought to any de- 
sired temperature and kept there, according to the demands of 
the process in question. 

3. Electricity offers us the cleanest heating agent imaginable;, 
so that we are enabled to avoid all deleterious influence which 
other heating agents have; for eleotric furnaces allow us to oper- 
ate in any atmosphere, and this prevents reactions taking place.- 



which may be caused by atmospheric elements, gases, or the 
products of combustion. 

4. The characteristics noted, in sections i and 3, allow the 
steel bath to be refined to any high degree. Sulphur particularly 
may be entirely eliminated, so that a high class finished product 
may be made from impure and cheap raw material. 

5. The electric furnace allows us to make crucible quality 
steel in large quantities (as mentioned in section 4), made from 
cheap raw material, and yet, at the same time, it turns out a com- 
pletely homogeneous product. This product has hitherto been 
possible only in the crucible furnace, where many separate 
crucibles are used charged with the purest and most expensive 
of raw materials. 

6. In many cases the product of electric furnaces shows cru- 
cible quality characteristics, even though the cheapest metal had 
been charged. This high quality cannot be achieved in any other 
type of furnace. The reason for this being that the heating 
agent does not in any way influence the charge and therefore the 
steel may stay in the electric furnace as long as deemed best, and 
held at any desirable temperature, meanwhile allowing the gases 
to escape. 

As these are the general principles which make the electric 
furnace valuable to the iron industry, it seems advisable to state 
the requirements which an ideal electric furnace would demand 
in order that the above advantages may be best attained. Par- 
ticularly as the number of different furnace designs are numerous. 

Surely everybody who is confronted with the question of 
installing an electric furnace, will see first that the installations 
shall cost the least amount of money, and second that the type 
used combines the greatest simplicity with the greatest safety 
during operation. 

The requirements, therefore, should be as follows: 

1. The ability to use any prevailing alternating current at 
any voltage and frequency. 

2. The avoidance of any sudden changes in the load. 

3. Ease of regulating the incoming current. 

4. High electrical efficiency. 


To which are added the following: 

5. A furnace of the tilting variety. 

6. Easily surveyed and accessible hearth. 

7. The electrical heating or any of its necessary auxiliaries 
must in no way influence the chemical composition of the steel 
or the slag. 

8. The ability to reach any desired uniform temperature in 
all parts of the bath, and at the same time avoiding any local 
under- or over-heating. 

9. The furnace should be as versatile in its application as 
possible. These requirements further stipulate the following: 

10. Equally advantageous, rapid, and inexpensive methods of 
removing all impurities contained in the charge, notably sulphur 
and phosphorus, and furthermore: 

11. The possibility of completely and easily removing any 
slag in the furnace, and of being quickly and easily able to 
renew it. 

12. Complete uniformity of the material in all parts of the 
molten metal and consequently a sufficient circulation in the bath. 

13. Avoidance of too much agitation in the bath, and there- 
fore providing an advantageous standing of the metal. 

14. The possibility of providing various furnace sizes, which 
would have to fit prevailing conditions. 

15. The highest possible thermal efficiency with all furnace 

16. The avoidance of all water cooling. 

17. The least possible refractory and initial cost and low- 
operating cost. 

It may be again remarked that the above requirements are 
those which would be expected of an ideal furnace. The furnaces 
discussed in the following chapters are those in practical use and 
therefore only partly fulfill the above requirements, some more 
and some less, so that the exactions made of an ideal furnace 
only serve as a normal estimate, with which the following various 
designs are compared. 

First of all, though, it seems necessary to dwell more in- 
timately upon the importance of several points. 


i. Of the furnace operation we required the use of any pre- 
vailing current. 

If this requirement were fulfilled it would enable any electric 
furnace to be connected to an existing central station, no matter 
if this were a city electric plant or the works' own isolated 
station. If' the electrical power was sufficient in either case, 
only the connection to the furnace installation and the latter 
itself would be necessary, so that the expense of a special genera- 
tor, which would only be ordered for the furnace itself, would be 

If, on the other hand, the case should present itself where the 
available power of an existing isolated plant was entirely in de- 
mand for other purposes, then in this case it would also be 
advantageous if any available current could be used for the 
electric furnace, so that the generator installation furnished for 
the electric furnace could at the same time and in any event be 
used as a reserve for the remaining generators ; or the generators 
would act as a mutual reserve, as well for the main generator 
installation as for furnace generators, which would then insure 
the best service conditions. 

If the consumer of electric current does not have to take 
into consideration the conditions existing in a distant central 
station when connecting to its lines, then such a connection also 
offers important advantages as it enables the existing central 
station current to be used. Furthermore the furnace installation 
in this case can easily be erected in a comparatively small place, 
besides saving the attendance for one's own power plant, or 
that required for a rotary transformer. This is entirely inde- 
pendent of the fact that small works are hardly able to generate 
power as cheaply as it can be sold by large central stations, 
excepting when high pressure internal combustion oil-engines are 

Accordingly, it would be desirable, of course, if direct or 
continuous current could be used for operating electric furnaces, 
in case a steel mill only possessed a direct current power plant. 
We, however, saw in the third chapter that on account of the 
chemical action of direct current, this does not appear suitable 


for operating electric furnaces, and as direct current can only be 
changed from a higher to a lower voltage, such as is used for 
arc furnaces, by means of expensive rotary converters, consisting 
of driving-motor and generator, and if the continuous current 
were to be used directly from a low voltage plant, the cost of 
the connecting wires and cables would be extraordinarily ex- 
pensive, as the distances are usually considerable; therefore, 
direct current is practically never used today for any electric 
furnaces in the iron industry. If, in spite of this, we see the 
assertion made here and there in advertising mediums, that a 
furnace may also be operated with direct current, then these 
assertions are to be approached with the greatest care, for when 
these are accurately tested, it will always be found that such 
allegations are misleading. 

It can accordingly be established that direct current does 
not come into play at all for operating electric furnaces. These 
latter may, however, be adjusted to any conditions which are 
offered by the modern alternating current station. 

It is well known that at present alternating current stations 
are built for three phase current, because the electrical conditions 
are especially favorable. When an electric furnace therefore is 
to be connected to an existing power plant, we shall no doubt, 
in the majority of cases, find that it is to be connected to a three 
phase plant. In this case a three phase furnace shall have a 
particular advantage which exactly fits into the conditions 
offered by an existing electric station. A two phase furnace has 
the same advantage as a three phase furnace, even though the 
former is to be connected to a three phase circuit, as three phase 
current may be changed to two phase by means of stationary 
transformers having the Scott connection. These transformers 
are necessary in such cases to regulate the power fed to the 
furnace, i.e., these regulating and phase changing transformers 
would serve the double purpose of simultaneously changing 
three to two phase current or vice versa, and regulate the current 
besides. Whereas a single phase furnace under these conditions 
would necessitate the installation of a rotary transformer, con- 
sisting of a three phase motor and a single phase generator, 


which would considerably increase both the initial and the 
current costs. But even though it would be necessary to install 
a new alternator to deliver current to the furnace, the three phase 
(or two phase) furnace has certain advantages. In this case it 
would be of considerable importance to obtain the least expensive 
electric plant consistent with economic operation. And it may 
be of determining importance here as a polyphase alternator costs 
about 25 to 33 per cent, less than a corresponding single phase 

If, on the other hand, single phase current only should be 
available, then the polyphase furnace would, of course, be more 
expensive, as the single phase current would then have to be 
changed to polyphase current by means of a rotary transformer. 
It appears, therefore, that the utilization of any existing single 
phase current would be of particular advantage. 

It seems that being able to use any voltage is of lesser im- 
portance. For as our requirements have limited us to the use 
of alternating currents, there no longer remains any noteworthy 
difficulty in changing or transforming a high central station 
voltage to a lower furnace voltage. For this change can be made 
very simply, and almost without loss, by means of stationary 
transformers, which only entail a comparatively small expense, 
and almost possess an unlimited life. 

Contrary to the foregoing, we find that it is of great im- 
portance to be able to use any existing frequency for the electric 
furnace. Unfortunately, this requirement is not yet completely 
fulfilled, practically, by any of the well-known furnace designs. 
Among others, the main reason is to be found in the power factor 
or cos <f> falling as the frequency rises. (See Chap. 4.) 

It can, therefore, only be established, (taking into considera- 
tion that only the practically attainable can be asked,) that an 
electric furnace should be operated with normal frequencies, 
meaning thereby 15, 25, 50 and 60 cycles. 

In order, however, to point out early, of what importance 
the periodicity of an alternator is regarding the cost, it may be 
mentioned for instance that the costs of a single phase alternator 
of 25 cycles and a similar one of equal capacity, but of only five 


cycles, will bear the ratio of i 12. These figures may perhaps 
best show, characteristically, the influence of abnormally low 

We now come to the second requirement, viz. : the avoidance 
of all sudden and untoward changes in the load. 

That such load changes and principally current fluctuations 
are of the greatest disadvantage to every electrical power plant, 
needs no explanation. It may only be remarked here that no 
city lighting and power plant would allow an electric furnace 
on its lines, which operated with heavy power fluctuations, with- 
out first interposing a rotary transformer with suitably heavy 
fly-wheels or other appurtenances which would be able to absorb 
these fluctuations and thus keep them away from the central 
station. This same requirement would also have to be met 
with in every other isolated plant, if any value is placed on its 
economical operation. 

With interposed rotary transformers, therefore, the power 
fluctuations would increase the initial cost. This also holds, 
provided the furnace is connected to a special generator. For 
it is evident that the generator must stand the greatest current 
fluctuations without injury, i.e., the generator must be built for 
much higher currents than if there were no irregular power surges. 
In other words, a generator required to operate a furnace, having 
current fluctuations, could operate a much larger furnace which 
was free from such fluctuations. To this must be added the 
fact that the generator's prime mover would run under much 
more unfavorable conditions, and with a much poorer efficiency, 
if the current surges are to be overcome, than if it only had to 
deliver the power uniformly or at a gradually changing rate. 
The pov/er delivered to an electric furnace, having power fluctua- 
tions, is similar to that taken by an electrically driven rolling 
mill or by an electric railway. 

In order to give an arithmetical example, it may be said that 
normally turbo-generators have a steam consumption of 7.5 kg. 
per kw.-hr. (16.5 Ibs. per kw.-hr.), whereas turbo-generators for 
railway service, with their required overload capacity, often have 
a steam consumption of 8.25 (18.15) and more up to 10 kg. per 


kw.-hr. (22 Ibs.). These figures about give a correct idea of the 
advantage which an electric furnace has whose operating force 
is free from fluctuations. This is quite apart from having a 
less expensive power plant which a smooth running furnace 
has. Furthermore, a power plant subject to having power 
fluctuations, is naturally liable to much greater wear than is 
occasioned by uniformly loaded machines. 

To the third point, viz.: the ease of regulating the electric 
furnace so as to give higher or lower temperatures, nothing 
more can be added. This is fulfilled as the furnace voltage 
can be easily changed, by suitable electrical apparatus, so 
that this requirement is fulfilled by all furnaces in the 
same way. 

Likewise the fourth point leaves nothing to be said regarding 
the requirement for a furnace with the highest possible efficiency. 
For, it is self-evident that a poor efficiency would entail a greater 
power absorption for the same work, and thereby the operating 
costs might be considerably increased. 

The remaining requirements refer mainly to metallurgical 
facts, which are discussed in detail in the second part of this 
book. That is why they are only given here just sufficiently to 
enable one to judge the different electric furnace designs. 

As a comparatively great number of charges are treated in an 
electric furnace, especially when operating with hot metal, 
nearly all the furnaces in practical operation today are made of 
the tilting variety. For this allows the teeming to be accom- 
plished with greater ease, and avoids much trouble caused by 
the giving away of the tapping hole. Consequently the demand 
for tilting furnaces today is a general one. 

In like manner there is recognized the demand for an easily 
surveyed and accessible hearth. For, every metallurgical 
operation will be placed in jeopardy without it. Therefore 
electric furnaces should have working doors placed at moderate 
heights above the bath, a little to one side, from which it should 
be possible to see the entire hearth. This may be required, for 
instance, in order to exactly determine the condition of the slag, 
or to be convinced when changing them, that the bath is really 


free therefrom before endeavoring to make a new slag. This is 
entirely independent of the fact that side doors are by far the 
most advantageous and convenient for charging slag. On 
account of the absence of an easily surveyed hearth, such resist- 
ance furnaces as described in the third chapter, having channels 
running to and fro, are absolutely to be discarded. 

It seems self-evident that we should expect an electric furnace 
to have its heat, or the necessary appliances required to give it, 
without influence on the chemical composition of the steel or 
slag. For it is just by these means that the electric furnace is 
to prove its superiority over the older gas-heating type. This 
point is, therefore, to be borne well in mind with every different 
furnace design. For suppose we assume that at any time 
during the metallurgical process, for instance, during the oxida- 
tion period, the electrical heating should in any way favor 
the oxidation, then this electrical heat effect would also be 
present at any other time, i.e., during the reducing period, and 
the furnace would then consequently be working at a disadvant- 
age. Thus the harm of these effects is often greater than the 
good they do, as they are also present when they are not 

Every metallurgist will concede that it is justifiable to expect 
an electric furnace to reach any desired temperature and still 
avoid any over- or under-heating. That primarily every practi- 
cally desired temperature must be attainable is evident, when 
we consider that the electric furnace must enable us to reach 
the most advantageous temperature for every stage of the 
metallurgical process. This requirement, therefore, falls to- 
gether with the one requiring an easy regulation of the incoming 
energy. With all this, it is of particular importance that the 
entire furnace contents be heated uniformly, so that over- and 
under-heating is not to be feared; it is much more likely that 
there would be an over-heating. The former of these is hardly 
likely to occur in case considerable heat is carried away by 
the water-cooled appliances in connection with the electrodes. 
Borchers, in his 1898 address before the "Verein deutscher 
Eisenhiittenleute," said: 


"As a matter of fact we need not fear that we cannot reach 
almost any temperature by electrical means for this or that 
purpose, for we shall have to place much greater weight on 
guarding against wastefulness on account of working with too 
high temperatures." 

The ninth point requires the electric furnace to be as versatile 
in its application as possible, and thereby possess the greatest 
adaptability in order to work it in conjunction with present or 
future processes. It goes without saying that it would be par- 
ticularly advantageous for the electric furnace, if it were possible 
to make in it the greatest variety of steel, equally well and 
economically, and of the same good quality. For even though 
one or the other object of making the steel may primarily be the 
absolutely determining factor, it is still to be noted that the 
electric furnace has a far-reaching application even today. 
However, there are at present still many new fields open to its 
product. So that even though it does not appear to be absolutely 
necessary, still by far in the most cases it would appear to be 
advantageous, provided a qualified electric furnace, or some 
chosen system, fits into the working program equally well for 
the reception of a new quality, as the previous material did. 

The further requirements from the loth to the i4th are self- 
evident, if the previous demands made upon the electric furnace 
are to be fulfilled. As the principal advantage of the electric 
furnace lies in the fact that it can turn out the highest quality 
steel from the cheapest raw material, it must consequently be 
easy to attain the removal of the impurities contained in the 
charge, provided the electric furnace economically permits what- 
ever refining there may be to do. First, we shall have to concern 
ourselves with the entire elimination of the phosphorus and 
sulphur; while removing the impurities which alloy themselves 
with the iron, (such as copper, for instance,) is also thus so far 
impossible in the electric furnace. If all the refining possible is 
to be carried out, it is absolutely necessary that the slag for re- 
moving the phosphorus, for instance, can be completely removed 
from the furnace. For otherwise, when the metallurgical proc- 
ess is continued for the removal of other impurities previously 


taken up by the slag, the phosphorus will again be taken up by 
the molten metal. The requirement of being able to completely 
remove slag from the furnace is covered, therefore, by doors 
enabling us to have an easily surveyed and accessible hearth. 

It seems just as self-evident that the impurities be removed 
from all parts of the bath, as it is necessary that all alloys added 
to it are absorbed equally by all parts of it. Otherwise an un- 
even material would result. On this account, therefore, a good 
electric furnace has to have an adequate circulation, which assures 
the greatest uniformity of material in all parts of the hearth. 
The desired agitation, however, must not exceed certain limits, 
as otherwise the advantage of the electric furnace would not be 
used which allows any slag solutions to be separated from the 
furnace contents. 

Finally, in order that the furnace can have a far-reaching 
application, it is necessary that the furnace be built of such sizes 
which seem to best fit present or future installations. This is 
to be kept in mind, for instance, when the furnace is to operate 
as an adjunct to a converter or an open hearth plant. In such 
cases, it is, of course, advantageous, if the furnace can receive a 
whole charge from a converter. It is such reasons as these that 
make it desirable to build furnaces of the largest capacity. 

The 1 5th requirement exacted a high thermal efficiency, and 
no explanation of this is necessary. However a few words may 
be said regarding the possible influence of using water cooling. 
First of all, it is evident that energy losses are caused by every 
cooling means, and water cooling aids this in the strongest degree, 
thus lowering the efficiency. Water cooling may become partic- 
ularly harmful when it is used in such manner as to considerably 
cool those wall parts which encircle the molten metal. For 
then the danger arises of the fluid iron assuming a certain tough 
fluidity, at these places, which makes it very hard to obtain a 
uniform composition of the entire furnace contents. Finally 
the employment of water cooling may easily cause dangerous 
explosions if the devices used are not very well protected. These 
would occur if the molten metal ever reached the water 


It only remains to mention the last requirements consisting 
of the lowest installation costs, likewise the lowest refractory 
cost, and thereby the lowest operating cost which brings together 
all the exactions which an ideal furnace has to fulfill. Un- 
fortunately, the complete attainment of this ideal has so far not 
been accomplished by actual practice, as evinced by electric 
furnace construction. This will be considered in the following 
chapters, where the constructions, as used, are compared with 
the stipulated requirements. We will find there, that every 
furnace design has certain advantages, but also certain disadvan- 
tages compared with every other electric furnace design. And it 
is this which makes the choice of a furnace thus far so difficult, 
for practical experience and the race in the open market have not 
yet perceptibly proved the superiority of one or another furnace 




IF the ends of two current carrying wires are brought together 
so that the current may flow, and if the two ends are then slightly 
separated, no interruption of the current will take place. But 
there will appear a small, highly luminous flame between the 
ends of the wires, which takes the place of the conductor at the 
point of interruption. With this, then, we have to deal with an 
entirely different property from that which the electric spark 
presents. The latter also represents a current transference 
through the air. But far higher voltages are necessary for the 
production of a spark than the arc calls for, an example of which 
we have just given above. In the latter, it is not the air which 
bridges the current, but the gases emanating from the metal of 
the wires between which the arc has been struck. The way the 
arc occurs then is as follows: 

At the instant when the ends of the two wires are separated, 
a rise of resistance of such magnitude appears at the point of 
separation, that, with the current flow, a corresponding and im- 
portant heating effect takes place. It is under this influence 
that the metal evaporates at the points of contact. If the 
separation should be increased, then the distance between the 
wire ends becomes so filled with metallic gases, that these now 
take up the current transference at the point of interruption. 
The metallic gases, however, are much poorer conductors than 
the metal itself. It follows then that the current in its path, 
from the end of one wire to the other, has to overcome consider- 
able resistance. The current flowing through this resistance 
gap generates such high temperatures, that more metal is gasified 
at the gap, in this way maintaining the arc. If no provisions 
have been made for hand or automatic regulation which keeps 
the distances between the wire ends constant, then the arc will 



rupture itself. This will happen as soon as the distance between 
the rigid wires is so large that the potential provided is no longer 
great enough to overcome the resistance of the arc. Should the 
arc be interrupted and it is desired to create it again, then the 
same ends of the wire must be brought together again, so that 
the arc may again be struck. 

It may be noticed, when striking an arc, that the metallic 
gases of the positive wire end or anode are carried away violently. 
This keeps the metallic gases together in a comparatively con- 
tracted area, thus making a definite path for the current. 

Even though the conducting metallic gas stream heated as a. 
resistance between the electrodes is absolutely necessary in 
order to maintain the arc, it can be interrupted by thrusting a 
cold body into the arc stream, although the maintenance of the 
arc is being upheld by an entirely permissible distance. When 
the arc is broken a decided cooling off then occurs at the point 
of interruption. This phenomenon is also to be considered with 
the operation of arc furnaces. 

From the above it is evident that every arc furnace furnishes 
that temperature which is required to gasify the conductors 
between which the arc is to be made. For the gasification of the 
conductor ends is the hypothesis upon which the maintenance 
of an arc rests. 

The best known arc formation is that which we see in the 
ordinary arc lamp. Here the arc is usually made between two 
carbon electrodes. 

The arcs in electric furnaces are made in a very similar way, 
for here carbon electrodes are also used to form the arc. As 
before said, this arc gives a very high temperature, in fact the 
highest which has so far been reached; for in the carbon we 
possess the most resistive conducting material, and this gasifies 
at about 3500 C. This gives us then the arc temperature with 
which iron and steel baths are heated in arc furnaces. 

Figs. 37 to 39 show the various possibilities which may be 
utilized for heating metal baths by the electric arc. In the 
schematically shown arrangement of Fig. 37, where the arc is 
formed directly between two carbon electrodes, we have the 



purest arc heating. The heating of the bath takes place by 
means of the radiating heat of the arc. The hearth is immedi- 
ately underneath the arc. These furnaces are generally known 
today as radiating furnaces. The best known arc furnace using 
this form of arc heating is the Stassano furnace, which will be 
discussed in detail in the next chapter. 

Figs. 38 and 39 show the main idea of two other heating 
possibilities when using the arc. These methods have the 
common characteristic of the hanging carbon electrode, which 
allows the arc to impinge itself directly against the metal. In 
both of these cases the metal bath is part of the electrical circuit, 
so that theoretically speaking we no longer have an exclusive 
arc heating. For the charge, composed of slag and metal, 

\ . .jHR.. JIR. / 


FIG. 37. 

FIG. 38. 

FIG. 39. 

naturally offers a certain resistance to the part of the current, 
by the overcoming of which heat is generated, no matter what 
the amount may be. Even though the resistance heating of 
the metal does not practically enter into the question at all, the 
designation of calling these furnaces combined arc and resistance 
furnaces is at least theoretically correct. Nevertheless these 
two furnaces have radical differences. The one shown by Fig. 
38 has the electrodes of all poles or phases above the bath, 
whereas with the furnace shown by Fig. 39, one pole is above, 
the other is constructed in a suitable position below the bath. 
The best known application of the former possibility is the 
Heroult furnace, whereas the equally well known Girod furnace 
embodies the second qualification. 

The question arises, where does the real heating take place, 
in furnaces as shown by the Figs. 38 and 39, where the arcs 
impinge directly against the bath? 


Although this question is discussed in detail in the chapter 
on the Girod furnace and an arithmetical example given, still 
the general answer to this may well be given here, which Borchers 
gave in 1905 in an address before the "Verein deutscher Eisen- 
hiittenleute." A translation of this follows: 

"When the electric current leaves the electrodes, a layer 
of air, gas or some vapor is formed between the electrode and 
the slag, which is a heat generating resistance in the circuit. 
This, therefore, gives us the possibility of arc heating. The 
bottom surface of the electrode encompasses about 1000 sq. cm. 
(155 sq. in.), at 3000 amperes. Thus we generate in every 
second a heat quantity of (Q = .246-1), or say 30 kilogram 
calories, in the small space between the electrode and the slag, 
even though we only assume 40 or 50 volts as the arc voltage. 
This amounts to over 100,000 calories given off hourly from the 
foot of the electrode. It follows that the slag layer is the second 
resistance between the electrode and the metal. The heat thus 
transformed is dependent on the thickness of the slag layer 
and on its constantly changing conductivity. If we take for this 
an additional drop of 10 volts, we add an additional 26,000 
calories, which is entirely independent of the small amount of 
heat appearing in the high conducting iron itself. The main heat 
therefore manifests itself in the space between the electrode and 
the slag. The foot of the electrode thereby has the gasifying 
temperature of carbon. A very considerable portion of the 
heat, therefore, enters the bath through radiation and through 
the carbon vapor, having over 3000 temperature, (C.) which is 
constantly thrown from the electrode onto the slag surface and 
is for the most part greedily absorbed by the oxygen in the slag." 

From the foregoing general characteristics of the combined 
arc and resistance furnaces, as they may be alluded to theoreti- 
cally, it follows that the heating of the metal bath takes place 
practically almost exclusively through the arc heating alone, so 
that the above furnaces are fully entitled to be simply referred 
to as arc furnaces, which is the case in practice. 


One of the most important parts of all arc furnaces are the 
electrodes, at the ends of which the arc is maintained, and which 
lead the current to the bath. 


A most resistive material is required for electric furnace 
electrodes in any event, and only carbon meets the requirements 
for those coming directly in contact with the bath, (if we omit 
for the moment the electrodes of iron or conductors of the second 
class,) as of all the metallic conducting materials, carbon alone 
stands the highest temperatures. It is, of course, to be con- 
sidered throughout that carbon is very liable to enter into reac- 
tions, especially at the temperatures found in electric furnaces, 
so that the metal bath must be protected by a layer of slag 
against an undesirable absorption of carbon, as is done for 
instance in the Heroult furnace for steel making. If this is 
done, carbon offers by far the most desirable material for arc 
furnace electrodes. 

These are made in specialty factories, or in case of very large 
electric furnace installations at their own works. They are 
made by hydraulic presses, being later on carefully dried and 
burned. Here one should strive to obtain a complete uni- 
formity of the mass, and the greatest mechanical solidity. 

Regarding the electric conductivity, it is to be noted that 
this varies greatly when using either carbon or the various sorts 
of amorphous carbon, charcoal, coke or soot. We obtain a 
higher conductivity, the more the finished electrode approaches 
the graphitic state, pure graphite electrodes giving the very highest 
conductivity obtainable. This item is dwelt upon later in detail. 

Moving parallel with the increase in the electrical conduc- 
tivity is the heat conductivity, so that when we have these 
favorable electrical conditions, i.e., when using graphite elec- 
trodes, we have the smallest Joule or i 2 r losses. To be sure, 
the largest thermal losses occur at the same time, because 
graphite electrodes, being good conductors, transmit large heat 
quantities from the inner furnace to the outside. 

We then have before us the interesting question concerning 
the most advantageous composition for the electrodes, i.e., 
rinding out how to gain their best efficiency. This question is 
of great importance, as the efficiency of the electrodes largely 
influences the total efficiency of arc furnaces. Before going into this 
question, however, we will preface it with a few general remarks. 


Next to the heat generated in the electrodes, current density 
has the greatest influence. That is, the number of amperes per 
unit of electrode cross-section, which of course accompanies the 
electrical conductivity. In accordance with a paper read before 
the "Verein deutscher Eisenhiittenleute," by Professor Borchers, 
in 1908, an electrode material having, the conductivity of arc- 
lamp carbons, commences to gasify its carbon when the current 
density is from 10 to 15 amperes per square millimetre (6500 to 
9750 amp. per square inch), whereas when the current is from 
.5 to i.o amperes per sq. millimetre, (325 to 650 per sq. inch), 
the temperature attained was from 500 to 600 C. 

Of course these current densities just mentioned do not 
occur in electrodes for electric furnaces. According to A. Helfen- 
stein, the electrodes of calcium carbide furnaces reach a red heat 
with only 9 to 10 amperes per square centimetre (58.0 to 64.5 
amps, per sq. inch). The considerably higher temperatures of 
carbon electrodes as used in practice in arc furnaces is explained 
by the electrodes not being heated by their ohmic resistance 
alone (r r loss), as they are heated besides this by the arc 
temperature at the electrode end. The following table 1 , taken 
from the book by Wilhelm Borchers, "The Electric Furnace," 
may show the current densities usually figured with: 

Electrode Diam. 

Carbon Cross-Section 
per Ampere 

Electrode Cross-Section 

Load in Amperes 
per Unit 
















12. 12 













30 to 40 

. 0465 to 


III. 2 

3-33 to 

215 to 



1 60 


15 -94 

60 to 90 

.093 to 



I . 66 to 

107 to 


i. ii 


1 The later (1912) type of electrodes used in the 15-ton Heroult furnace at 
South Chicago are 20 inches in diameter. As they carry from 9000 to 1200 
amperes per phase, the current density is as low as 30 to 38 amperes per sq. 
inch (see also page 307). 


It must be seen from the table that the load per unit of cross- 
section decreases as the electrode cross-section increases, the 
essential reason being that the manufacture of electrodes of the 
best quality becomes more difficult as their cross-section increases. 

It is also evident, that it is harder to make a completely even 
mass in a large electrode cross-section, than in a small cross- 
section. It is likewise much easier to obtain an even annealing 
for thin electrode rods, than for thick rods. Finally the gasifica- 
tion of part of the binding material of electrodes is much more 
uniformly and completely accomplished in small cross- sections, 
than is possible in large cross-sections, in which it is almost im- 
possible to avoid irregularities. If these facts illustrate the 
decrease of the permissible current density with increasing 
electrode cross-sections, and if it appears that the use of too large 
cross-sections is not advisable, we find that the considerable 
weight of the carbon electrodes is also forbidding; besides there 
is irregular solidity with growing cross-sections. 

Owing to this, it has been found preferable, sometimes, to 
build up large electrodes of several smaller ones and thus avoid 
one large electrode block. See Figs. 6oa and 6ob. 

Thus we can use in these smaller electrodes the higher per- 
missible current densities, and attain a smaller total electrode 
cross-section, which consequently give the much desired lower 
thermal losses. With all this, we stand anew before the question 
of what is the best division between the electrical and thermal 
losses in the electrodes, i.e., how shall their best efficiency be 
attained? This theme has been extensively discussed in 1909 
and 1911 in the Electrochemical and Metallurgical Industry, 
latterly called Metallurgical and Chemical , Engineering. The 
principle articles are by C. A. Hansen and Carl Hering. Even 
though these dissertations could not solve the question of the 
best electrode dimensions completely, still, the results are so 
important that they are presented here in condensed form. 

As before mentioned there are two kinds of losses in the 
electrodes : 

i. Losses through Joule heat, i.e., those in consequence of the 
electric current flowing. 


2. Losses through heat conduction, i.e., those occasioned by 
the' electrode (being a good heat conductor) leading the heat 
from the inner furnace to the outside. 

How complicated these conditions become by the cooperation 
of these two losses is evidenced by one of Hansen's tests, for he 
obtained the astounding result, when the ohmic resistance and 
the current density were increased to such an extent that the 
Joule losses doubled still the total losses remained the same. 

Of what importance the clearing up of these conditions is, 
is very evident, when we hear that according to Hansen the 
losses in a furnace operating with 500 Kw. can easily be 15 
per cent, of the total energy input. This would be continually 
75 Kw., or with a current cost of 3^ cent per Kw.-hour, the 
electrode losses would cost about 56 cents hourly. 

When the electrodes are incorrectly dimensioned, it may 
happen that thermal losses increase to such an extent, that it 
is no longer possible to keep the whole bath molten. Then only 
just that part which is directly beneath the arc will stay molten, 
while the remainder will remain solid, owing to the heat transfer- 
ence occasioned by the extravagant dimensioning of the electrodes. 

These examples already show that a saving in the electrode 
losses may, under certain circumstances, be the deciding factor 
for the economic working of the electrode furnace, especially if 
the price of current be high, while in other cases large sums of 
money could be saved, if we succeeded in approaching as 
nearly as possible the best theoretical electrode dimensions. 

In order to become acquainted with the conditions governing 
the least losses, Hansen made parallel tests with graphite and 
carbon electrodes which gave the following results. 

The efficiency of graphite electrodes grows with increasing 
length and increasing current densities. It is, however, im- 
possible to force the current density above certain limits, as the 
electrodes then taken on temperatures that are too high, which 
might easily destroy the surrounding brickwork. 

With ordinary carbon electrodes an increasing length causes 
a decrease in the efficiency, whereas the Joule effect becomes 
much larger than the thermal losses. 


The experimentally ascertained conditions of Hansen are, of 
course, only true between certain limits. It is evident that by 
continuing to increase the length of graphite electrodes, up to 
a certain point, a condition would soon result where the 
losses are a minimum. If this point is exceeded then the Joule 
effect would increase more rapidly than the heat losses would 
decrease, and this would result in an increase of the total losses. 

Similarly the minimum losses would be exceeded if the current 
density were increased beyond its best value. 

These reflections led Her ing to determine the most favorable 
electrode dimensions theoretically. Even though these deter- 
minations do not always give the greatest consideration to the 
conditions in actual practice, and the results may only be partly 
used in practice, still they give such interesting disclosures, re- 
garding occurring conditions, that they are for this reason worthy 
of note, and will be given a little later on. 

Now next it is evident, that under any conditions and inde- 
pendent of material, an increase in the electrode cross-section 
increases the heat conducting losses, simultaneously, though 
decreasing the electrical losses. On the other hand a lengthening 
of the electrode, namely on the inside of the insulating brickwork, 
causes a decrease of the thermal and an increase of the electrical 
losses. When both cases are extreme the losses will be infinitely 

It is a fact, however, that the Joule heat as well as the heat 
carried off through conduction are both generated by electricity. 
Thus the object is to bring the total losses down to a minimum. 

For this it is quite necessary to know accurate values of heat 
conductivity and specific resistance for every -electrode material. 
Furthermore, there should be accurate results on the indepen- 
dence of these values of the temperature. Unfortunately such 
results are almost wholly missing. To all this must be added 
the fact that all electrodes are manufactured articles, which 
are not capable of being produced of complete uniformity, and 
the constants of these, (of proved material,) are greatly dependent 
on the chosen cross-section. 

On account of all these reasons, these figures, ascertained 


from the present meagre tests, cannot lay any claim to general 
validity. Therefore, the following values must also be cau- 
tiously used and should be considered primarily as approximate 

Hering, in his computations, now entirely neglects the heat 
quantity which is given off by the electrode, and which is absorbed 
by its surrounding brickwork. He assumes that the electrode 
is insulated throughout its entire length, so that the heat is only 
conducted away by the end of the electrode, which is on the out- 
side of the furnace and there usually cooled with water. He 
further assumes that the electrode has the exact same cross- 
section throughout its entire length, and that the change of 
conductivity with temperature follows a straight line. All these 
are assumptions which are not borne out by the facts, but are, 
however, necessary in order to make the conditions for theory 
and practice more distinguishable. 

With the assumptions as made, the conditions may be 
visually shown by Figs. 40 and 41. In Fig. 40, E E represents 



FIG. 40. 

FIG. 41. 

an electrode, which is surrounded about its periphery with a 
complete heat insulator, so that only the ends remain free, which 
are for instance kept cool by means of water-cooling. If we now 
send a comparatively heavy current through the electrode, this 
will heat the latter strongly at the middle point H, until an 
equalizing condition occurs. As soon as this is reached, the 
entire Joule heat will be carried off at the cooled electrode ends, 
while no heat flow will occur at H any more. 

If we now cut the insulated electrode at H, in order to utilize 
both parts as electrodes for an electric furnace, as it is schemati- 
cally shown by Fig. 41, there will be no change in the situation, 
provided the furnace has the same temperature which it formerly 


had at H, assuming of course that the same current strength as 
before now flows through the electrodes. Under these con- 
ditions also, there will be no heat loss from the furnace interior, 
through the electrodes to the outside of the furnace. 

The condition given herewith is the ideal one, so that we 
may have only the minimum electrode losses, with, of course, 
the previously made assumptions. 

Provided the assumptions have the limitations as originally 
laid down, the losses will be equal to 

where Qi equals that heat loss, which would be carried from the 
furnace to the outside by the electrode, if no electric current 
were flowing, and Q 2 equals that heat quantity which is solely 
and alone generated by the current overcoming the electrode 
resistance = i" r. 

Consequently Qi = c k r y 

here c = 4.18, a constant, which is used for converting gram 
calories into watts. 

k = the mean conductivity in gram calories per second by i 

cm. length and i sq. cm. cross-section, with the tem- 

perature difference appearing between the hot and the 

cold electrode ends. 

r = temperature difference between the hot and cold electrode 


/ = the length of the electrode in centimetres. 

6*> -2 

2 = r r = ? Pl 

where r denotes the total resistance of the electrodes. 

Pi = the mean specific resistance per cubic centimetre at 
the occurring temperature difference. 

/ = length in centimetres. 

q = cross-section in square centimetres. 


We therefore obtain the total losses, which are carried away 
from the cold electrode, as 

that is: 

The total energy losses carried off at the cool electrode end, are 
equal to the sum of the heat losses, which would occur if no current 
flowed through the electrode, plus half of the heat lost by means of 
the Joule effect. 

The losses are a minimum when the pure heat losses are equal 
to one-half those caused by the Joule effect, i.e., when 

4.18 k T -j- = i 2 pi 

v 2 Q 

In this case the total losses are equal to the Joule heat losses or 
i" r, and hence no heat would be carried from the furnace by 

From the equation for the minimum losses, 

it follows that 

If this result is then substituted for -y- in the general equation 


for the total losses, we have: 

In order to attain the minimum losses we have the require- 

Q min = 2.89 * VTT^i 

The equation for Q m i n shows that the minimum losses are de- 
termined by the material constants k and pi, the temperature 
differences between the hot and cold electrode ends, and the 
current strength. It is independent of the absolute dimensions 
of the electrodes, for of these it is only required to maintain a 
definite relation between the cross-section and the length in 

accordance with the equation for y 


If we substitute for the equation for Q min the specific con- 

ductivity per cubic centimetre, K = , in place of the specific 


resistance, we obtain 

Q min = 2.8 9 i^ r 

This equation shows that the least losses are fixed for a certain 
definite temperature on account of the relation between the heat 
conductivity and the electrical conductivity. In accordance 
with this the best material for the electrodes is that which has the 
least heat conductivity and the highest possible electrical conductivity. 
From the equation for the minimum losses, it follows that an 
increase of the temperature difference between the hot and 
cold electrode ends only influences the losses in proportion to the 
square root of these differences. 

If we again consider the equation, 

we see, that with a given material, a given current strength, and 
a given temperature difference, the electrode losses would remain 
the same for entirely different cross-sections, provided the pro- 
portion between the cross-section and the length remained un- 

From this we now learn : // it be desired to save on electrode 
material when having a minimum of losses, then the electrode is to 
be made as short as possible. Generally the electrode length is 
primarily determined by the practical demands of the furnace 
operation, so that a certain minimum distance of electrode 
length cannot be exceeded. The length of .electrode, therefore, 
having been determined, the cross-section can be calculated by 

using the formula -y 

It is well to mention here that it can be assumed that all these 
calculations only retain their full correctness, provided the 
electrodes are protected by insulation throughout their whole 
length. If this assumption has to be dropped, which is necessary 
for practical purposes, then an increase of the cross-section in 


the same proportion to the length, causes a decided increase in 
the electrode surface and with it naturally an increase of the heat 

From the derived formulas, it is evident that the current 
strength influences the size of the losses. Consequently, it would 
be requisite to have the smallest possible current at high voltages. 
Unfortunately, this demand cannot be fulfilled, without leaving 
the total efficiency of the furnace out of consideration, for it is 
always well to keep in mind that the electrode losses considerably 
affect the furnace efficiency, but are not the sole factors that 
carry weight with it. This point is discussed further on. 

In place of the current strength in the formula 

we can insert the current density A = , 
and obtain 

/= 2 .8 9 - I -=JU 
A \ Pl 

This formula produces a combination showing the best con- 
ditions between cross-section and length together with a current 
density fit for use. This seems advantageous, because by over- 
stepping the permissible current density limits, it is very easy 
to endanger the furnace operation. On page 84 mention has 
been made of these tests by Hansen. However, this formula 
also has the disadvantage, that it determines the electrode length 
arithmetically, which is not fully determinable for practical 
reasons. And the value of this derived formula, practically only 
consists in bringing forth a clear idea of the conditions of an 
ideal case. It should be the ambition of every furnace designer 
to come as near to this as possible. 

In order to be able to utilize these rules and references, it 
would be necessary to have useful constants for the different 
conductivities of different electrode materials. Unfortunately 
there has been a great lack of these up to the present. Even 
though a few values are given hereafter, it must be observed that 
they have reference only to a certain definite material, which just 


happened to be used for these determinations, and that products 
from other factories would give results deviating from these, 
more or less. Nevertheless, the figures comparing the graphite 
and carbon electrodes may be regarded as typical, and can con- 
sequently be used in practise for electrode designs. 

Hansen gives the following figures for temperature differences 
up to 3000 Centigrade: 

Material. Pi k 

Graphite 1 000812 .16 

Carbon 2 00183 .016 

The proportion between the electrical resistance of carbon and 
graphite is as 2.2$ : i, whereas the heat conductivity of graphite 
is ten times as great as that of carbon. 

Relative to the current densities Hansen believes it safe to 
figure with the following values: 3 For graphite, 150 amps, per 
square inch. This equals 4.3 sq. mm. per amp. or 23.25 amp. 
per sq. cm. For carbon, 50 amp. per square inch. This equals 
13 sq. mm. per amp., or 7.75 amp. per sq. cm. Substituting these 

values in the ratio -y- in accordance with the equation : 
7 = -345 ' ^ 

* K T 

for instance for 20000 amp., and 3000 C., we have for graphite. 

q \ .000812 

and for carbon, 

q .000183 

7 := -345 X 2000 ^ OJ6 x 3QOO - 42.57 

i.e., for equal electrode lengths (which would be required for the 
same furnace) of graphite as well as for carbon, a carbon electrode 

would have to have = 4.73 times the cross-section of a graph- 
ite electrode. 

1 Specific resistance ohms per inch cube = .000320. 

2 Specific resistance ohms per inch cube = .000721. 

8 Compare the values dependent on the cross-section as given on page 82. 
See also page 81. 


After the relation for is given, we can either assume the 

required electrode length as given (on account of the practical 
furnace requirements), and thereafter determine the cross- 
section, which could then be regulated by permissible current 
densities. We could, however, also figure from a given current 
density as a basis, and from the cross-section thus determined, 
calculate the electrode length, which would then have to have its 
practical applicability proved. 

Should we choose the latter method, we may calculate the 
cross-section based on a certain current density deemed per- 
missible, and based on Hansen's values, for instance, for the 
electrode cross-sections of this material. The example cited 
was for 20,000 amperes. We then have: 

graphite = - = 860 sq. cm. 

/ 20000 . , 

(= ~^~ = 133 sq. inches) 

on account of the proportion, therefore, of y = 9, we obtain a 

length of 95 cm., or 37.4 inches. 

If we assume that this length is satisfactory to the furnace 
operation, then this same length will, of course, have to be kept 
for the carbon electrode, and in case a minimum of losses is also 
desired here, we would have for the carbon electrode cross- 

section (based on the calculated relation of y = 42.57) 

q carbon = 4044 sq. cm. 

(= 1591 sq. inches). 

From this, with 20000 amps., we have a current density of 

A = ^Z~ = 4<9 amp ' per sq ' cm ' 


. 20000 N 

(= -- 12.5 amp. per sq. in.), 

which would show that according to the values given on page 82, 
these are sufficiently high, so that an enlargement of the cross- 
section would recommend itself, and perhaps a simultaneous 



increase in the electrode length, in order to stay as close as 
possible to the minimum losses. 

Besides this it is interesting to become acquainted with the 
losses as they appear in the given example, either when using 
graphite or carbon for the electrodes. The equation for the 
minimum losses was : 

Q min = 2.8g 

By substituting the values for graphite, we obtain Q min = 36 
KW and for carbon Q min = 17 KW., i.e., assuming that the given 
constants are correct, the losses for graphite would be about 
twice as large as those for carbon. 

This condition, however, only holds good, when the electrodes 
are heat insulated for their entire length as previously mentioned. 

In accordance with the values heretofore cited on page 82, 
for the usual current density values for electric arc furnace 
carbon electrodes, it seems that the figure of 7.75 amps, per 
square centimetre (50 amps, per square inch), which Hansen 
gives, is extraordinarily high. It is therefore not advisable to 
use these figures, which gives much too short electrodes for 
practical furnace constructions, as the example showed. It is 
better to use those values given on page 82, which simultaneously 
take into consideration the influence of the electrode cross- 
section enlargement. 

The figures given in the following tables are from tests made 
by Hansen and published by him. On the one hand for graphite 
electrodes made by the International Acheson Graphite Co., 
and on the other for carbon electrodes made by the National 
Carbon Co.; these may show the influence of the cross-section 
enlargement on the material constants even a little better. 

Diameter or 

pi = Resistance 
ohms per cm. cube 

Diameter or 

PZ = Resistance 
ohms per in. cube 

5.08 cm. diam. 
7.62 cm. diam. 

. 00092 to . 00093 
.00103 to .00109 

.00096 to .00101 
. 00084 to . 00085 

2 inches diam. 
3 inches diam. 
4 in. X4in. 
6 in. x6 in. 

. 000362 to . 000366 
. 000406 to . 000429 
.000378 to .000397 
.000331 to .000335 


These measurements were made at a temperature of 25 C. 
With increasing temperature the resistance of graphite falls 
as is well known. According to Hansen, this is as follows: 
At 25 Centigrade 100% 









6 5 % 


For carbon electrodes Hansen found the following values de- 
pending on the cross-section: 


Size of 

pi = Resistance 
in ohms per ccm. cube 

Size of 

P2 = Resistance 
in ohms per inch cube 

15.24x15. 24 

. 00856 
.00594 to -007I 
.014 to .0254 

4 in. x 4 in. 
6 in. x 6 in. 
Sin. x Sin. 
iSin. x iSin. 

. 00234 to 00279 

.00551 to .0100 

Referring to the last of these values, it is well to note that this 
test was made on an electrode delivered 4 years ago, and it is 
possible that better results have been attained since then, for 
large electrodes. 1 

Hansen also made some investigations with carbon electrodes 
in order to determine the influence of temperature. He found 
that, with an increasing temperature, the carbon continually 
proceeded to graphitize, so that after the electrodes had cooled 
down, the original figures for the specific resistance no longer 
held true, but were, instead, much better. 

The following table shows how the specific resistance of the 
cold carbon electrode falls, in case the electrode has been pre- 
viously heated to the temperature shown in the table: 

1 In 1912 the National Carbon Co. state that their "Steel furnace elec- 
trodes have a resistance of about .0025 to .0030 ohms per inch cube." 


Resistance in the cold condition 100% 

After heating up to 1200 C 91 . 6% 

" 1600 C 87-% 

" 2000 C 77-6% 

tt (t tt O'/~* r rt-f 

2400 C 65 . 9% 

a (( <( o o /~ rrf 

2800 C 50-9% 

" 3500 C 22.4% 

Here the last figure approaches that which would be obtained 
with graphite electrodes under the same conditions. 

Besides, Hansen gives as an average figure of many tests made 
with commercial carbon electrodes when heated to 1200 C., a 
resistance value equal to 60 per cent, of that measured in the 
cold state. 

After the electrodes have once been in operation, the uni- 
formity of the material constants disappear in all parts of the 
cross-section or the length, owing to the uneven heating of the 
carbon throughout its entire length. On this account Hansen 
takes the practical resistance at 1200 C., at only 40 per cent, of 
its cold figure. 

As for the rest, we again point to the figures which were used 
in the arithmetical example on page 76. 

The remarks regarding the best dimensioning of the electrodes, 
have a certain practical significance, and that is why they have 
been discussed here. It is well to be warned, though, that too 
great stress be not placed on these theoretical opinions. 

It is to be noted that the derived formulae are only strictly 
accurate for such cases, where the electrode is protected from 
heat losses between its hot and cold ends and that this case 
never appears in practise. It is further to be observed, that the 
operation of our arc furnaces necessitates a shortening of the elec- 
trodes, and consequently considerable electrode lengths appear, 
which are not taken into consideration in the formula, because 
they lie outside of the water cooling. Furthermore, the formula 
are not the only measure for the losses which actually appear in 
arc furnaces, irrespective of the restrictions just made. Besides 
the pure radiating losses, there are for instance the contact losses, 
where the current carrying copper conductor clamps onto the 


electrode. And above all it is a noticeable fact, that on the one 
hand, the size of the cross-section is of the greatest influence on 
the efficiency of the furnace, while on the other we see that the 
electrode length is primarily settled by practical considerations 
accompanying the furnace operation. 

We see, therefore, that just this relation between cross- 
section and length of electrode, which is of striking importance 
in accordance with the formula for the minimum losses, cannot be 
freely determined according to the arithmetical values. After 
all, the benefit of the calculation for the minimum losses lies in 
the fact that it allows us to ascertain the heat dimensions which 
lie between the given limits of practical requirements, that we 
may come as near them as possible. 

Turning again toward the practical side of the electrode 
question, we find an interesting work of Hansen's, which deals 
with the burning away of the electrode or -electrode consumption. 

This question is, of course, of equal importance, as the striv- 
ing after the least electrical losses, or a high efficiency; for this 
point is of considerable influence on the operating costs. 

The consumption of electrodes may occur: 

1. In the worst case when the electrode breaks; 

2. By the arc formation which causes a gasifying of the carbon 

3. By oxidation. 

Those under the first heading, which are by far the most 
unpleasant, seldom or never occur today, as long as the cross- 
section and lengths used are not too large. Too large cross- 
sections are always to be avoided, so that if high currents cannot 
be avoided, it is better to use graphite in place of carbon elec- 
trodes. Furthermore, it is to be observed that a new electrode 
must not be placed in the hot furnace in its cold state, as small 
particles are easily liable to crack off, on account of the great 
prevailing temperature differences. It is therefore commendable 
to heat the electrodes slightly before placing them in use. 

The losses under the second heading are self-evident and 
unavoidable, so that nothing remains to be said about them. 


On the contrary a much greater interest manifests itself in 
the electrode consumption on account of the oxidation. 

Moissan found that amorphous carbon commences to oxidize 
at as low a temperature as 375 to 490 C., whereas graphite first 
begins to oxidize at temperatures of 665 to 690 C. These values 
though were observed with powdered material and not with solid 
rods. 1 Finally, Collins, FitzGerald, and Johnson maintain that 
graphite possesses a greater resistivity against oxidation than 
carbon does. 

Contrary to this, Hansen observed that the losses with 
graphite electrodes are greater than those with carbon electrodes . 
In making these tests, graphite rods of the International Acheson 
Graphite Co., and carbon rods of the National Carbon Co. were 
used. The reason for the higher consumption, when using 
graphite rods, as given by Hansen, is that at temperatures of 
1300 to 1400 C., the graphite particles cracking off are so large, 
that some of them could be picked up unconsumed. This 
phenomenon disappeared when the heating occurred in carbonic 
acid gas, which proves that the cracking off, of the electrode 
particles, when using graphite, leads us back to the oxidizing 

This investigation shows that it is impossible to accurately 
determine in advance just what the electrode consumption will 
be. For this is so dependent on all oxidizing influences, that 
even the tight or less tight closing of the working doors, or the 
piercing of the electrodes through the furnace roof, or the working 
in a more or less reducing atmosphere, may cause considerable 
changes in the electrode consumption. 

Oxidizing losses not only affect the electrode consumption, 
but the power consumption of the furnace as well, i.e., the 
efficiency. This is evident from the following tests. 

Hansen operated a small Heroult furnace of 150 kg. (330 
Ibs.) capacity, with graphite electrodes of io.i6x 10.16 sq. cm. 
(16 sq. inches) cross-section and 106. cm. (40 inches) long. 

1 These values, however, seem quite reliable since we find data published 
from electrode manufacturers which give the temperature of oxidation in air 
at 640 C., and 500 C., respectively, for graphite and carbon. 


With this arrangement he succeeded in melting a 150 kg. charge 
with 150 kw.-hrs. 

Later on a similar furnace was operated, but for 300 kg. 
(660 Ibs.), having the same electrodes. It was established here 
through various tests, that the power consumption of the larger 
furnace had the ratio of 1.2 to i compared to the smaller furnace, 
even though a larger furnace usually has comparatively smaller 
thermal losses than a smaller furnace. After graphite electrodes 
of 15.24 x 15.24 cm. sq. (36 sq. inches), and 101.6 cm. (40 inches), 
long were used, the larger furnace gave a somewhat better power 
consumption than the smaller one. 

The tests further showed that the power consumption rose, 
as soon as the electrode (the end toward the molten metal) be- 
came more and more pointed under the oxidizing influences. 
The difference in power consumption when working with the 
full cross-section compared to the operation with a pointed one 
was as much as 30 per cent. 

This test, as well as others made with various electrode cross- 
sections in the 300 Kg. (660 Ib.) trial furnace, show that a larger 
cross-section causes a decrease in the losses. This may be 
primarily caused by the fact that a larger cross-section permits 
a more favorable dissemination of energy throughout the whole 
charge, and furthermore, because the full and larger electrode 
cross-section acts as an umbrella, which considerably lessens 
the heat radiation toward the furnace roof. The umbrella 
action of the electrode also has the additional advantage of 
keeping the roof from deteriorating too rapidly. This, however, 
changes as soon as the electrode takes on its pointed form. 

Hansen established that a more or less strong sharpening to a 
point of the electrode occurs in all arc furnaces, under the oxidizing 
influence which takes place during the working period. The trials 
carried out to protect the electrodes by suitable coverings of car- 
borundum, water-glass, etc., against the oxidation, have not been 
successful, for it has not been possible to make the covering 
durable with the prevalent temperature differences, occurring 
during the furnace operation. We, therefore, have to figure with 
a, certain burning away of all electrodes, owing to the oxidation. 


Aside from the three reasons, which have so far been given 
to determine the electrode consumption, there must still be 
mentioned the additional loss caused by the stub ends. The 
length of this stub end depends largely on the distance between 
the molten metal and the furnace roof. Recently newer methods 
have been devised which now render it possible to attach the 
electrode remainders to the new electrodes, thus assuring a most 
complete use of the electrode material. This is gone into further 
in the chapter on the Heroult furnace. 

During the discussion of the electrode conditions, we have 
often compared the graphite with the carbon electrodes. Is 
therefore one recommended above the other? To this question 
this reply may be given: Graphite electrodes mainly have the 
advantage of greater resistivity, and greater mechanical firmness. 
This advantage, though, must be purchased at a far higher 
price, compared to carbon electrodes. Large electrode surfaces 
tend to save energy, and consequently it is better to work with 
low current densities. For the graphite electrode loses its im- 
portance, i. e., its high electrical conductivity, whereas its 
disadvantage of a high heat conductivity falls heavily in the 
balance, so that the graphite electrode always has a lesser 
efficiency than the carbon electrode (see page 94). 

From all this it is apparent, that one would at first endeavor 
to utilize carbon electrodes, at least as long as these can still 
be made of good quality and at the desired cross-sections. It 
is only with the largest furnaces, where the cross-sections would 
become so large, that uncertainties would enter the operation, 
through breakages, for instance, that one would be willing to 
pocket the disadvantages of the graphite electrode, in order to 
gain the important advantage of definite and sure operating 


Previously when discussing electrode conditions it was 
always assumed some water cooling would be arranged at the 
place where the electrode leaves the furnace roof, by means of 
which it would be possible, to lower the temperature of the 


electrode as much as 100 or 200 C. It was also shown that the 
electrode material may occasion considerable losses on account 
of the oxidation, and this gives us the first reason which forces 
'the application of electrode cooling upon us. 

If sufficient cooling was provided, so that the electrode, 
where it issues from the furnace, is not cooled below the tem- 
perature, where the oxidation begins, then the unavoidable 
oxidation in the circulating atmosphere would considerably 
reduce the cross-section. This would be followed by an increase 
in the electrical resistance, hence a stronger heating up of the 
cross-section already weakened, and therefore an increasing 
temperature with increasing consumption, so that in the shortest 
space of time a change of electrodes would be required. 

If an intensive water cooling is already unavoidable, this 
will simultaneously act protectingly on the uniformity of the 
furnace operation. Thus the contact arrangements which 
connect the copper conductors to the carbon electrodes are kept 
from being destroyed. Supposing we assume that the electrode, 
even outside the furnace, has a comparatively high temperature 
as well, then there would be such an increase in the heating of 
the contact pieces, tha.t their hold on the electrodes would be 
loosened, and with other designs they would burst, so that in 
both cases the furnace operation would fail, on account of a 
break in the electrode contacts, quite irrespective of any damage 
done by the flames shooting through the roof, where the elec- 
trodes enter. 

The water-cooling device is also responsible for the long life 
to-day of the arc furnace contact clamps. At the same time, it 
fulfils a third and very important purpose. It was shown in 
Chapter II that all refractory materials used in electric furnace 
construction are conductors of the second class, and as such 
obtain higher conductivities with increasing temperatures. This 
also holds for the brickwork between which the electrodes of arc 
furnaces lie. It is apparent that these roof bricks become more 
and more conducting with increasing temperatures, whereas they 
can be regarded practically as non-conductors with low or even 
moderate temperatures. In order to avoid a strong oxidation 


of the electrodes, and to attain the best possible thermal effi- 
ciency, it is necessary to have the closest fit where the electrodes 
protrude through the furnace roof. Thus, it is immediately 
apparent, that when the roof refractories are little resistant, i.e., 
when their temperature is high, then the small spaces between 
the electrodes and the surrounding roof bricks are easily bridged 
over with tiny arcs, which in turn cause currents to flow through 
the refractory material from one electrode to the other. 

The current flowing through the brickwork will be higher, 
as the voltage increases between the different electrodes, as the 
distance between the electrodes becomes less and the temperature 
of the brickwork between the electrodes rises. That these 
currents flowing through the refractory material may be of great 
importance is shown in an article by Coussergues after seeing a 
Stassano furnace. In a one-ton furnace, when the arc was 
interrupted and the voltage was 1 20, there was still a current of 
300 amperes flowing through the brickwork from electrode to 
electrode. It is to be noted here, that the entrance of the 
electrodes to the furnace is provided with water-cooling contri- 
vances. If an attempt were made in such a case as this, to do 
away with the water-cooling, then the temperature of the brick- 
work in the neighborhood of the electrodes would rise consider- 
ably, the resistance between electrode and electrode would 
thereby further decrease, and still stronger currents would 
traverse the brickwork. The result would be a considerable 
increase in the energy consumption, while a strong heating 
ensues at the wrong place, and at the same time there would 
be a quick destruction of the very highly heated roof due to the 
current flowing. 

In accordance with the foregoing, it is established that the 
utilization of water cooling with arc furnaces offers important 
operating advantages, even though there is, of course, a certain 
heat loss on that account, which is unavoidable up to certain 
limits. Aside from this there is still, under some circumstances, 
a small electrical loss, which may appear when currents from the 
electrode find their way to the cooling chambers, and are thence 
grounded by the water. 



In discussing the arc it was shown that it can only be main- 
tained, provided a certain distance between the electrode and 
the bath is not exceeded, as otherwise the arc will be interrupted. 
It is therefore necessary to watch the length of the arc. This is 
easily accomplished with the aid of a voltmeter or an ammeter. 
The electrodes are then regulated in accordance with readings 
of the controlling instruments. 

Even though a manually operated regulation of the electrodes 
is possible, as, for instance, with the Stassano furnace, we find 
the equipment with automatic regulation, as used in the arc 
furnaces of Heroult and Girod to-day, have several advantages. 

In both cases, i.e., either hand or automatic regulation, this 
is accomplished with the aid of gears, which are driven by an 
electric-motor in order to handle them faster and more accurately. 
In accordance with the indications on the measuring instruments 
the motor is started either to the right or left by throwing a 
double-throw switch, which either raises or lowers the electrode. 

The Thury regulator, invented in 1898, is used almost ex- 
clusively for this automatic regulation. It is made by Ateliers 
H. Cuenod, A.G., at Chatelaine near Geneva, Switzerland. 

The principal part of a Thury regulator is an electro-magnetic 
scale, which is balanced when the current and voltage conditions 
are normal. When deviations occur in the normal circuit 
conditions, they throw the lever out of balance. These scales 
are used then to throw a switch to either one side or the other, 
so that the current for the driving motor enters it either from 
one or the other side, thus bringing about the corresponding 
motion of the electrode. 

The switching mechanism of the Thury regulator consists of 
a small constantly running auxiliary motor, which moves a lever 
back and forth. This lever engages a suitable pawl and ratchet 
mechanism so arranged, that when the electro-magnetic scale 
is not in balance, it releases one of two pawls which then catches 
the teeth of a wheel, and causes it to revolve in one direction or 
another, by the aid of the pendulum motion of the lever, carrying 
the pawls, at the same time the shaft of this latter wheel carries 



the switch, which operates the driving motor. Fig. 42 shows 
this mechanism. 

The electro-magnetic scales which bring about the desired 
regulation, are built for either direct or alternating current. It 
operates as a volt, ampere, watt or ohm meter and is provided 

FIG. 42. 

with a regulating resistance, which allows the operating conditions 
of the furnace to be changed at will. The double- throw switch, 
which controls the driving motor is either single or double pole. 
The current is broken between an adjustable copper piece and 
a block of carbon of generous dimensions, so as to equalize the 
burning away of the contacts, or to lengthen the time of contact. 


If several furnaces are to be automatically regulated, only 
one driving motor is required for all regulators. The apparatus 
are then mounted on a switchboard, which also carries the con- 
trol instruments such as volt and ammeters. 

The regulators are also provided with a manually operated 
switch, which cuts out the automatic regulation, so that hand 
regulation may be resorted to. The motor drive for the pulleys 

FIG. 43. 

is kept, however, which is often very desirable when using hand 

Fig. 43 shows an assembled view of all the apparatus, which 
is necessary for an automatic regulation of the electrodes. There 
is a train of gears, of which a pinion and rack either raises or 
lowers the electrodes. The electrodes may also be suitably set 
by the aid of cables or chains. The weight of the electrode and 
its appurtenances may be partly equalized by a suitable counter- 


The foregoing has briefly discussed the more or less common 
phenomena and appliances of arc furnaces, and hereafter some 
of the various designs of arc furnaces will be gone into. The 
most important things of arc heating may again be briefly stated 

In all arc furnaces the heating of the bath is brought about 
practically exclusively by the arc itself. There are always 
temperatures of about 3500 C., occasioned by the arc. Even 
with a moderate heating this temperature cannot be avoided. 

Borchers, in his 1908 address before the "Verein deutscher 
Eisenhiittenleute," said about this: 

"In arc furnaces there may be many arcs, the arcs may also 
be brought in more or less great distances from the bath, in order 
to bring this to a temperature of less than 3500 C.; but 3500 C. 
is always generated at some restricted places, and we must 
operate downwards from this temperature." 



IT was shown in Chapter VI, that among the better known 
electric furnaces, the Stassano furnace is the only one which is 
exclusively operated by arc heating. We may, therefore, also 
refer to it as a radiating furnace. 

It was Stassano's original ambition to build an electric blast 
or shaft furnace. His object was primarily to use profitably the 
rich ore fields of Italy, where native coal is scarce. 

His first patent, issued in 1898, in England, is based on the 
following claim: "The utilization of caloric energy of the 
voltaic arc for primary determining the reduction of oxide of 
iron and the metals to be combined therewith and afterwards 
melting the metallic masses reduced, for the purpose of obtaining 
in a fluid state the product desired, all substantially as set forth." 

The furnace which Stassano suggested for this trial is shown 
by Fig. 44 in plan and vertical cross-section. Without describ- 
ing the first design of this furnace at length, it may be briefly 
said, that Stassano laid great stress on the point that no air was 
permitted to enter the furnace. With a furnace of this kind 
Stassano made his first tests in Rome. With 1800 amperes at 
50 volts he succeeded in producing 30 Kg. (66 Ibs.), of metal in 
one hour. 

As a result of these trials, a furnace plant for the direct 
reduction of iron ores was erected at Darfo, in Lombardy, Italy. 

Despite several changes in the construction of his furnace, 
Stassano, though keeping his method of heating, was not able to 
give any permanent life to his electric shaft furnace. When the 
Canadian Commission made their observation trip in 1904 the 
installation at Darfo was no longer in existence. 

In the meantime, Stassano had forsaken the original design 
of the shaft-like construction, and instead built a hearth furnace 

1 06 



with an inclined bottom as shown in Figs. 45 and 46. This 
furnace in which the ore was charged underneath the arcs, in- 

FIG. 44. 

FIG. 45. 

FIG. 46. 

stead of at the top, as in his shaft-like furnace, was intended for 
both the reduction of iron ores to pig iron, and the refining of 


pig iron to steel. As the figure shows, the furnace was meant 
to have three pairs of electrodes, which could all be used at once, 
or singly, for striking the arc, so that the temperature of the 
furnace could be regulated. 

But even with this suggestion for a pure arc furnace Stassano 

FIG. 47. 

could not achieve success. He was however able to make a 
new furnace installation at Turin, Italy, in which he first used a 
rotating furnace. This furnace was patented in all industrial 
countries, and dated about the year 1902. As this furnace is 
in use to some extent today, it will be discussed in detail, showing 
as it does the best known furnace with purely arc heating. 



Figs. 47 and 48 show the furnace in vertical and horizontal 
cross-section. It is very evident from the claim of Stassano's 
patent that he laid particular stress on the motion of the molten 
metal in the furnace. 

As the drawings show (Figs. 47 and 48), the rotary arrange- 
ment of the furnace necessitates a vertical cylinder. The shell 
of the furnace is constructed of sheet iron, and is connected at 
the lower part, near the bottom, with a strong ring-shaped carrier, 

FIG. 48. 

which in turn rests on rollers. The motion is usually transmitted 
by gears driven by an electric motor. At the middle of the 
furnace bottom, axial to the direction of the furnace, we find 
the current and water supply. The current is brought in by 
means of brushes and slip rings, such as are found on any poly- 
phase motor. The water cooling, which is brought from the 
fixed to the movable part by suitable means, is needed for two 
purposes with this furnace. First, it serves as cooling water 
for the electrodes, and again as the water under pressure for the 


electrode regulation. The figures show that the furnace hearth 
is covered with a double sort of roof, which is not readily re- 
movable with this type of furnace. This arrangement allows 
the heat protecting qualities of the brickwork to be utilized to a 
great extent, and, as a matter of fact, this method is said to give 
an extraordinary heat insulation, which can even be bettered 
by inserting layers of lime or sand. 

Fig. 47 shows an outlet in the upper part of the melting 
chamber, allowing a free escape of the gases which are generated 
during the reaction. This outlet pipe is surrounded with a sand 
filled covering, into which it dips at its lower end. This pipe 
does not take part in rotation of the furnace, but is kept in place 
by suitable means. 

The gas removing system again betrays the ambition of 
Stassano to smelt ore, and serves to protect the furnace com- 
pletely from the entrance of outside air. As a matter of fact, 
however, this furnace did not give satisfactory results for smelting 
ores directly. On this account this furnace is today used only 
for the working up of scrap or for refining hot charges. 

In such cases, therefore, this gas flue falls away, as in the in- 
stallation of duplicate Stassano furnaces at the Bonner Maschin- 
enfabrik, Bonn, Germany. The hearth here has also been given 
a hexagonal shape, whereas Fig. 48 still shows the round form 
as used by Stassano. 

The bottom of the furnace consists of Magnesite brick, 1 
as does also the double form of furnace roof. The insulating 
layers of furnace refractories are partly comprised of tamped in 
material. The furnace is provided with a door for watching the 
metallurgical work, for charging the metal, adding the slagging 
materials and rabbling it off, for taking samples, etc. Besides 
this the furnace bottom is supplied with a tap, through which 
the finished material flows. 

The most essential and most important furnace part is, of 
course, the arrangement of the electrodes. As the furnace may be 
built as well for single phase as for three phase, it would have 

1 According to Stahl und Risen, page 1066, 1910, the side walls and bottom 
are said to have lately consisted of tamped in dolomite. 



two or three electrodes, as the case may be. These pierce the 
furnace walls as is plainly shown in Figs. 47 and 48 and form 
an arc or arcs in the middle of the furnace which heat the bath. 

Stassano laid great stress on the design of bringing the elec- 
trodes through the furnace. The electrodes enter the furnace by 
first piercing double walled cylindrical chambers. There is a 
circulation of water in the space surrounded by both walls, in 
order to keep the temperature of the outer electrode portion 
down. There is a regulating cylinder over each cooling cylinder, 
the former aiding the setting of the electrodes to any desired 
point. The piston-rod is connected at its outer end by means 
of a sliding guide rod with the one end of another rod, which 
carries the electrode itself at the other end, which latter end is 
in the cooling cylinder. In order to better show this arrange- 

FIG. 49. 

ment, the whole design of this electrode regulating apparatus is 
shown in Fig. 49 on a larger scale. 

The regulating of the electrodes is accomplished without 
any automatic regulating apparatus, but is accomplished manu- 
ally by the aid of the hydraulic cylinder. Any common water 
pressure of 4 or 5 atmospheres (60 or 75 Ibs.) can be used, so 
that no special water pumps are needed for the electrode regula- 
tion. The current carrying parts are naturally easily and well 
insulated electrically from the furnace shell, as short circuits 
would otherwise occur through the furnace walls. Stassano 
did not look with favor upon any automatic regulating appa- 
ratus for his electrodes, and Osann who studied the operation 
of the Stassano furnace in detail gave the following reasons in a 
report in Stahl und Eisen, 1908: "An automatic regulating 


arrangement would be complicated in any event and would not 
be advisable if for no other reason than this alone because the 
electrodes are withdrawn while charging; besides this, an elec- 
trode breaking off now and then is not precluded, and this frag- 
ment must be removed quickly. This is simply and quickly 
accomplished by calling to the man who watches the three 
ammeters, and operates the three corresponding levers which 
control the hydraulic cylinders for the electrodes. The electrodes 
can be used up until the remaining stump only protrudes .1 m. 
(4 inches). Then they are changed, and this change takes only 
from 3 to 5 minutes, all of which, I have personally assured 

We now come to the behavior of the furnace during its 
operation. As already mentioned, the electrodes are withdrawn 
when the furnace is being charged. When the furnace is about 
two- thirds charged, the electrodes are brought together, to 
again form arcs and are then regulated by watching the needles 
of the ammeters. The charging of the furnace with scrap takes 
about 15 minutes for a i-ton furnace, and the setting of the 
electrodes thereafter, takes about two minutes. As soon as the 
first scrap is melted down, 'the remainder is charged on top of 
it, but this time without withdrawing the electrodes, i.e., without 
any interruption of the current taking place and working with 
the utmost speed, so as to avoid all radiating losses. The slag- 
forming materials are charged in the usual way, and the dephos- 
phorizing slag is likewise removed after the dephosphorizing 
period is over, similarly to the practise with any other electric 
furnace. In order to easily remove the slag, the furnace is 
turned far enough, so that it may be conveniently removed 
through the door. This is possible as the furnace axis has a defi- 
nite angle of about 7 from the vertical, so that the door assumes 
different positions toward the bath surface, during the turning. 

If after this general characterization of the Stassano furnace, 
we turn to one of its definite examples, we find that the duplicate 
furnaces at Bonn of i-ton size are the best, being one of the later 
Stassano furnace installations. 

These i-ton furnaces of 250-!!? are built for three-phase 


current; no volts is needed to operate them. The current is 
supplied from a distant central station at an incoming voltage 
of 5200. This voltage can, of course, not be used directly in the 
Stassano furnace, and is consequently transformed in a separate 
transformer, removed from the furnace, and stepped down to the 
aforesaid no volts. During the normal operating condition, 
the furnace takes from 1000 to noo amperes at 105 to no volts, 
and this current is held as steady as possible throughout the en- 
tire operation. The Stassano furnace having a very good power 
factor, (as high as .9 to .95 per cent.,) the energy consumption 
for this i-ton three-phase furnace is 1.73 X noo X no X .95 = 
198.86 Kw., or say, 200 Kw. 

It is necessary to have a man watch the electrical conditions. 
He regulates the arcing distances of the electrodes, by means 
of the levers controlling the hydraulic cylinders, and watches 
the ammeters, one of which is in each phase. The rotating 
motion of this Stassano furnace in Bonn is transmitted by means 
of a 5-HP motor to a tight and loose pulley, connected by a 
shaft to gears, one of which is a part of the furnace. The electrode 
diameters of all Stassano furnaces are kept down as much as possi- 
ble, so that the work is carried on with comparatively high cur- 
rent densities. In furnaces up to 500 HP, electrode diameters of 
80 mm. (3.2 inches) are used. According to an article by Cous- 
sergues in the Revue de Metallurgie, this diameter is also used 
in larger furnaces up to 1000 HP. In this case, however, the 
electrodes are doubled in number. 

Accordingly, for the 25O-HP furnace at Bonn, for instance, 
which takes noo amperes with its 80 mm. (3.2 in.) diameter 
electrodes, whose cross-section is 5024 square mm. (7.78 sq. in.) 
corresponding to 

i ioo /i 100 141 amps.X 

= .22 amperes per square millimetre (j^ -- per sq in j 

or 22 amperes per square centimetre. 

With a 5oo-HP furnace having the same electrode cross- 
section and about twice the current, the current density would 
rise to 


2200 /22oo 282. amps.X 

= 44 ampere per square millimetre^ = per sq J 

or 44 amperes per square centimetre. 

This, therefore, gives current densities which still substanti- 
ally exceed those given by Hansen, as mentioned on page 91, 
even though these values had to be designated as being high 
enough. We have also then with Stassano furnaces to figure 
with a substantial heat generation in the electrodes. A short 
example may show this. 

New electrodes for a 250-!!? furnace have a length of 1.5 
metres (59 inches). As the electrodes wear off during the 
operation, we may figure with an average length of i metre, 
(39! inches). If we insert besides this the operating value 
for the resistance of carbon per cubic centimetre, as given by 
Hansen and shown on page 76 to be, 

pi = .00183 ohms per cubic centimetre. 

we obtain the resistance of the electrode as being: 


r = pi where / and q are in centimetres and square centi- 
metres respectively. 
Consequently : 

r = .00183 - - = .0036 ohm. 

The drop in voltage in the electrode of a 250-HP furnace hence is 
e ir 

= noo X .0036 = 3. 96 -say 4 volts. 
The energy transformed into heat per electrode is consequently 

A = ie watts = noo X 4 = 4400 watts, 

or in all 3 X 4400 = 13200 watts. That is, with a total energy 
absorption of 200 kw. for the furnace, there is 6.5 per cent, lost 
through Joule losses (r r) in the electrodes alone. 

Besides the transformation of electrical energy into heat in 
the electrodes as just described, several interesting phenomena 
will be found in the Stassano furnace as shown below. 

First regarding the length of the arc, with Stassano furnaces 
with voltages of no up to a maximum of 150 volts, this distance 


at first is about 10 cm. (4 inches) from electrode to electrode. 
During the run, however, the arc distance increases up to a length 
of 30 cm. (about 12 inches). This considerable lengthening of 
the arc is partly accounted for on the one hand by the high 
temperature of the furnace atmosphere, and on the other hand 
through the gasification of the electrode ends caused by the arcs 
between them. 'It is to be noticed that the arc sags toward the 
bath. This phenomenon can only be regarded as favorable to 
the heating of the metal bath. 

We, therefore, find with the Stassano furnace, an increasing 
lengthening of the arc, as the temperature of the furnace atmos- 
phere increases. The risk must, therefore, be run of having 
the arc break and making it anew, when charging the furnace 
with cold material which cuts the arc. On this account, there- 
fore, particular care should be exercised when charging the 
furnace, entirely independent of the horizontal arrangement of 
the electrode rods. 

If, notwithstanding this care, the arc should still break, then 
the rise of the furnace temperature is interrupted until the arc 
is again established. Still there wou'd be no complete interrup- 
tion of the energy absorption. Thus, according to Coussergues, 
when visiting the Stassano furnace at Bonn, the arc was inter- 
rupted, yet 300 amperes per phase at 120 volts were still taken 
up by the furnace, which is about one- third of the total energy. 

This energy absorption with an interrupted arc is only then 
possible, if the refractories are heated to redness. For the energy 
absorption is dependent upon small arcs establishing themselves 
between the refractories and the electrode, which carry the 
current from the electrode to the magnesite bricks, after the 
latter have become conductors of the second class, due to the 
high temperature, and may therefore be regarded as heating 
resistances between the electrodes (see page 17). 

Finally attention may be drawn to the capability of Stassano 
furnaces to be heated up electrically, since the charge is com- 
pletely independent of the arc formation. In this way the 
furnace is also kept up to temperature during any shut-downs. 
This is accomplished by heating up for a quarter of an hour with 


the arc, followed by a current interruption for three-quarters of 
an hour. 

The above states the specific characteristics of the Stassano 
furnace. We now come to the comparison of the Stassano furnace 
with the ideal electric furnace, for which the requirements were 
laid down in Chapter V. Without entering into a discussion of 
the purely metallurgical questions which are gone into in detail 
in Part II of this book, we may say the following: 

The first requirement stipulated that the electric furnace 
was to be capable of being operated with any prevailing alternating 
current at any voltage and periodicity. 

This requirement is met by the Stassano furnace better than 
by any other of the well-known arc furnaces, for Stassano furnaces 
are built for single-phase as well as for three-phase current. At 
the same time any prevailing periodicity may be used. Opposite 
this, the necessity of transforming the voltage to that required 
by the furnace, only plays a secondary part, for this transforming 
takes place in comparatively inexpensive stationary transformers, 
which hardly call forth any particular vigilance, considering 
their great operating safety. 

The second requirement, viz.: the avoidance of sudden 
power fluctuations is better filled by the Stassano furnace than 
by any other furnace, especially when melting down cold stock. 
For aside from the interruptions during the charging period, as 
already mentioned, the sustaining of the arc is in noway influenced 
by the melting process, so that if the attendant regulating the 
electrodes is sufficiently attentive, sudden current fluctuations 
should be reduced to a minimum. These conditions permit 
Stassano furnaces to be connected directly to the line without 
the interposition of costly regulating apparatus. 

We now come to the third point in which an easy regulation 
of the current is demanded. This requirement may also be re- 
garded as being fulfilled, as voltage regulation, simultaneously 
causes a regulation of the energy supplied to the furnace, which 
is entirely independent of such energy regulation which is pro- 
vided by different settings of the electrodes. It was mentioned 
on page in that Stassano avoided every automatic regulation of 


the electrodes with his furnaces, which would still offer several 
advantages. These reasons are referred to again at this time. 

The requirement under 4, viz.: a high electrical efficiency, 
does not seem to be so completely fulfilled. We have already 
seen that the high current densities in the electrodes lead to im- 
portant heat losses, and it does not seem therefore that it is 
possible to avoid considerable losses. This is even accentuated 
by the intensive water cooling of the electrodes. Unfortunately, 
figures regarding these actual losses are nowhere to be found, 
still they cannot be unimportant, as the example on page 114 
indicates. Besides the electrode and cooling losses there are 
the transformer losses, for changing the voltage to the desired 
amount, for which about 3 per cent, of the total energy may be 

The fifth point, viz.: the tilting arrangement, which Stassano 
replaced with a turning one, no doubt gives his furnace certain 
advantages; still, compared to the tilting device, his solution 
can hardly be regarded as a particularly happy one. The turning 
or rotating structure requires a really complicated mechanism. 
As a proof of this it is only necessary to refer to the water supply 
for the electrode regulation and to the electrode cooling. En- 
tirely aside from this, it hardly seems advantageous to have a 
tapping hole, instead of pouring over the lip, when teeming, 
especially when heats follow each other quickly, as is usually the 
case when treating hot metal. 

Even though the requirement of an easily surveyed hearth 
seems to be completely fulfilled, it is yet to be observed, that the 
almost horizontal arrangement of the electrodes makes the ful- 
filment of the seventh requirement so much harder. For the 
breakable electrodes with their comparatively small cross- 
sections are liable to crack off when roughly handled, so that 
the metallurgical operations in the furnace entail great attention 
and not a little dexterity. Besides this, the Stassano furnace 
would have the advantage of influencing the charge the least 
with its arc heating, in case electrode breakages could be avoided 
with certainty, as the carbon vapor from the electrodes is not 
directly against the molten metal. Relative to the avoidance 


of every under or over heating of the metal, it must be said that 
the influence of the arc heating as employed by Stassano, i.e., by 
use of the radiation, is the mildest way in which arc heating can 
be used at all, as the direct influence of a heating agency of 3500 
C., on the metal, is avoided. 

Without discussing the purely metallurgical demands, the 
fulfilment or non-fulfilment of which can be readily seen by the 
construction of the furnace, we find that the requirement of a 
sufficient but not too strong a circulation of the bath is fulfilled by the 
rotary arrangement of the furnace. No other mechanical cir- 
culation appears in the Stassano furnace as it is built today, and 
it seems, therefore, that if any security is desired for a complete 
uniformity of the material in its several layers, it is not 
possible to dispense with the mechanical bath circulation. And 
these necessary mechanisms must always be designated as being 
very complicated (for any such metallurgical apparatus as this), 
no matter how ingeniously the design may have been carried out. 

Besides the many sided applications of this furnace, it would 
seem desirable if they could be built of any possible size. The 
proof of this is, however, yet to be established. For even though 
Stassano furnaces of 5-ton size were operated by Stassano him- 
self, at the plant in his charge in Turin, the plant unfortunately 
has been temporarily shut down. It may, therefore, at present 
only be regarded as proven that the Stassano furnace of 600 to 
1000 Kg. (5/8 to i ton), as it is operated at Bonn for melting up 
scrap for steel castings, succeeded in giving good results. The 
furnace does not seem suitable for larger sizes, as the sensitive 
devices permissible, at any rate, with small furnaces, while easy 
to watch, are hardly applicable with large furnaces. The high 
current density, with which 5-ton furnaces are to be operated, 
also seems unadaptable, while with still larger furnaces where 
the doubling of the electrode number would be encountered, 
difficulties could be expected from the simultaneous manual 
regulation of six electrodes. 

Larger furnaces would have longer and consequently more 
breakable electrodes, which would otherwise need much room 
during the furnace's rotation. Finally, the easy working of the 




furnace becomes almost impossible with the six horizontal 
electrodes over- topping the bath. All these reasons make it 
appear that the Stassano furnace in its present customary form 
is only useful for small capacities. 

The requirement of a good thermal efficiency is acceptably 
fulfilled. For even though the Stassano is the electric furnace 
where the metal is heated most indirectly, but where the atmos- 
phere directly above the bath is heated the strongest, and though 
it is not possible to avoid a considerable heat loss when the 
furnace door is opened, still it is well to note that the heat in- 
sulation with the Stassano furnace is extraordinarily well carried 
out, and that consequently it is possible to attain satisfactory 
power consumption figures for melting a ton of steel. And these 
vary between 800 and 1000 Kw. hours per ton of steel for melting 
cold stock for making steel castings. 

According to Osann (Stahl und Eisen, 1908, p. 660), we 
find that he begins with a cost of 62 cents for electrodes at the 
Bonn furnace and $2.75 for refractories per ton of steel, so that 
we cannot speak here of exactly low refractory costs, which could, 
however, be considerably reduced by using dolomite bottoms and 
side walls (Stahl u. Eisen, 1910, p. 1060). 

The installation costs for a i-ton furnace are given by Osann, 
inclusive of switchboard and foundation, at $8,750. This does 
not, however, say that the cost of the necessary transformer is 
included in this price. In Bonn the voltage is stepped down 
from 5200 to no volts. On the other hand, it may be said that 
at Bonn they were enabled to connect to an existing central 
station, so that in case such connection is not possible, the 
installation cost would be increased by an amount equal to the 
cost of an isolated plant (250 HP for a i-ton furnace). 

Fig. 50 shows a Stassano furnace from which the general 
arrangement is evident. Regarding the sale which these furnaces 
have had, reference is given to the list in Chapter XV. The 
giving of licenses for Stassano furnaces is made by the Bonner 
Maschinenfabrik und Eisengieszerei Fr. Monkemoller & Co., 
Bonn on the Rhein, Germany. 


HEROULT had already earned great merit in the development 
of electro-metallurgy, on account of his electric furnace for the 
production of aluminum. He was the first to discover how to 
build an arc furnace for refining iron, having vertical electrodes 
pointing directly at the bath. Before this these furnaces had 
the objection, that the iron bath greedily absorbed the carbon 
from the immersed electrodes. On July 4, 1900, Heroult made 
the suggestion (see German patent No. 139904), that to avoid 
the absorption of carbon by the metal bath, the slag used to 
refine the metal should be inserted between the bath and the 

According to the patent description the electrodes are to be 
so far separated from each other and are to dip so little into the 
slag, that, on the one hand, the resistance between the electrodes 
within the layer of slag, shall be great enough to force the current 
from the one electrode through the slag lying directly beneath 
it to the metal, and from the metal again through the same layer 
of slag to the other electrode, and that there shall be otherwise 
no connection between either electrode, and the metal. Further, 
according to the patent description, the striking of arcs between 
the electrodes and the metal bath into which the electrodes 
project, is not precluded, or is it necessary. Regulating the 
distance between the electrodes and the metal bath, however, is 
the important part. This must be accomplished in such a way 
that the slag layer between the electrodes and the metal bath 
remains hotter and more conductive during the entire refining 
period, than the layer of slag between the electrodes, because 
only in this way will the current take the path as prescribed 



After this general characterization of the Heroult furnace, 
and before entering into details regarding its construction and 

FIG. 51. FIG. 52. 

operation, we will give a short survey of the development of this 

According to the Electrochemical and Metallurgical Industry, 
1909, p. 261, Heroult, in his first efforts in building an electric 

FIG. 53- 

FIG. 54. 

furnace, leaned narrowly toward his type of aluminum furnace. 
In this furnace, as is well known, one pole consists of a hanging 



carbon electrode, while the other pole was made by the furnace 
hearth itself. For this purpose the hearth was made of carbon. 
When it was necessary, however, to obtain a material with the 
lowest possible carbon content, this style of furnace could no 
longer be used, as the carbon of the hearth bottom was greedily 
absorbed by the molten metal. 

On that account Heroult next made tests with a furnace for 
the production of low carbon ferro chromium. The bottom of 
this furnace consisted of chromite bricks in the middle of which 
a carbon block was inserted which then acted as the bottom 
electrode. With this method Heroult hoped that a part of the 

FIG. 55. 

carbon block would be absorbed by the molten metal and that 
the molten mass would continue to force its way down -absorbing 
carbon as it went, until the exterior radiation of the molten 
metal would cause it to freeze on the carbon block. Heroult 
hoped to keep this condition constant, so that there would be an 
interposition of the frozen metal between the bottom carbon 
electrode and the bath, which would at the same time prevent 
any carbon absorption by the bath. 

But the tests as carried out did not fulfil his hopes, and so 
after further trials there was produced the Heroult furnace as 
we know it today. This has been characteristically shown by 


the above examples, taken as they are, first of all, from the 
patent records. Furnaces of this kind were first put in trial in 
Froges and La Praz, France. 

The first Heroult furnace in Germany was installed by the 
firm of Richard Lindenberg of Remscheid in 1905, and put in 
operation in February, 1906. Since then the furnace has come 
into extensive use, thanks to its simple design, and thanks to a 
thorough knowledge of the metallurgical operations, which have 
been thoroughly investigated at Remscheid. 

Coming now to the furnace itself, we may say the following: 
Of all the arc furnaces the Heroult furnace resembles most nearly 
a tilting open hearth furnace. It consists of a steel plate shell 
of nearly rectangular form which has a rounded bottom. Fast- 
ened on to this are curved tracks, which permit the furnace to 
run in channels. The furnace is tilted by means of a hydraulic 
cylinder. The whole design of the furnace may be seen by 
consulting the Figs. 51 to 55 inclusive. 

The lining of the furnace consists of fire-bricks H, which are 
laid directly against the steel plate shell, and on which dolomite 
is tamped. The roof is removable. It therefore consists of a 
wrought iron frame, lined with fire-brick H, the former also 
having convenient screw eyes so that the whole may be trans- 
ported readily. The hearth may be easily inspected and the 
furnace may be easily operated during the charging period, as 
the furnace has three doors, one in the middle in front and one 
at each side. 

The arched roof of the furnace is pierced by two or three elec- 
trodes. Copper cooling chambers are placed at the piercing points 
(not shown in the illustrations), which keep the carbon elec- 
trodes outside of the furnace at permissible temperature limits 
(as discussed in Chapter VI), and simultaneously cool the brick 
work at these points. Each electrode hangs from a right-angled 
support R, which is movable in a vertical direction at the furnace. 
This support, therefore, carries a rack, which is moved by a 
motor-driven pinion. The use of these small motors in this 
design permits a mechanical regulation of the electrode positions. 
In Remscheid these small regulating motors are of the single 



phase loo-volt type. These motors operate automatically or 
by hand, according to whether a higher or a lower position of 
the electrodes is called for. Naturally the electrode clamps are 
insulated in an improved manner from the furnace casing. 

Regarding the development of the automatic regulating 
apparatus of the Thury regulator, this was described at length 
in Chapter VI, pages 102-104. The electro-magnetic scales men- 
tioned there are connected as a voltmeter to the Heroult furnace, 
as shown in Fig. 56 by the dotted lines. The voltmeter is de- 
signated by m in whose place we can imagine the electro-magnetic 
scales. Two scales are provided as each electrode is regulated 
separately. The scales are influenced by the voltage which lies 

FIG. 56. 

between the head of the electrode and the bath, which receives 
the main current. In order to obtain this voltage an iron rod 
is. embedded in the furnace bottom, which in turn is connected 
to the remaining terminals of the magnetic scales. These scales 
are set so that a difference of two volts from the normal will 
start the regulator and keep it as near constant as possible. 

The design of the furnace is such that either hot or cold 
charges may be treated. With cold charges, however, very 
heavy fluctuations of the current cannot be avoided, until the 
whole charge is melted down. The reason being that it is much 
harder to melt down a stone cold charge in an arc furnace and 


maintain the arc, than it is to treat hot material. This is be- 
cause an arc furnace always operates better at a high temperature, 
under the influence of which carbon evaporates. Furthermore, 
the appearance associated with the so-called over regulation 
causes the electrodes to become unruly, when, for instance, the 
heavy current fluctuations occur by the arc rupturing and 
establishing itself again. During the melting of a cold charge, 
continuous fluctuations therefore follow, and these continue 
until the charge has become molten. During the time of these 
heavy current fluctuations, that is while melting the metal, the 
automatic regulation is replaced by hand regulation with the 
Heroult furnace also. But the series connection of the carbon 

S-^-t- Automatic -Regulating 

!l 12 l3 ! 4 'D G 

FIG. 57. 

electrodes has a negative influence on the electrical conditions 
even with perfectly fluid metal. To illustrate this, Fig. 57 shows 
several current curves, as they were recorded by an arc furnace 
with series connected electrodes. These curves are reproduced 
from an article by Englehardt in the Zeitschrift des Osterreiches- 
chen Ingenieur- und Architekten-Vereins, during 1909. 

In order that no misunderstanding may arise regarding the 
heating method of the Heroult furnace, it is well to especially 
mention at this time, that Heroult had soon to realize that he 
must employ an arc to make his furnace operate even though it 
deviated from the furnace operation of his patent description. 


In this it was not precluded nor was it necessary that arcs should 
be struck between the electrode and the bath. Hence to-day the 
furnace voltages are chosen so high, that the electrodes are set 
at about 45 cm. (18 in.), above the steel bath. With this setting 
it is possible to obviate a carburization of the bath when the 
slag is interposed, and this is solely caused by the heating action 
of the arc, (having a length as mentioned above,) heating the 
metal to the desired temperature. 

If the electrodes in the Heroult furnace were dipped into 
the slag, so that no arc exists, then the furnace would be of the 
pure resistance variety. Should we now calculate the resistance 
conditions in such a circuit, we shall immediately find, that, under 
these conditions, practically the whole energy would be changed 
into heat in the electrodes, without heating the bath materially 
at all. This is apparent when we compare the resistances of the 
two carbon or graphite electrodes connected in series, with their 
comparatively small cross-section and very great length and 
their high specific resistance, with the resistance of the slag layer 
and the bath with their very large cross-sections and very short 
lengths and the very low specific resistances (at least as far as 
the bath is concerned). These conditions have been clearly 
recognized by the representatives of the Heroult furnaces. We 
quote from Prof. Eichhoff of Charlottenburg, the technical 
adviser of Lichtenberg of Remscheid, his article appearing in 
Stahl und Eisen, 1909, p. 843, as follows: 

"It is impossible to heat an arc furnace for steel, by utilizing 
the heat generated by the resistance of the thin slag layer or the 
large cross-section of the bath. These resistances only furnish 
a few per cent, of the heat necessary in the furnace," and again, 

" Obtaining heat by the rising temperature of the slag with 
its decreasing resistance, or by utilizing the resistance of the 
bath, has never been achieved, simply because the slag layer is 
too thin, and the cross-section of the steel bath too large. Such 
a view therefore is a fable, which I oppose from the start." 

This description should suffice to give a perfectly clear 
picture of the workings of a Heroult furnace, in which then 
practically the entire heating is obtained from the heat of the arc. 


If we return for the moment to the furnace design we observe 
the following: 

Heroult furnaces are usually built for single phase currents 
of from 25 to 33 cycles. The Heroult furnace, for instance, at 
La Praz operates on 33 cycle current at no volts. The charge 
is about 2^ tons. At this rate the furnace consumes about 
4000 amperes. With very large 'furnaces 1 Heroult uses 3 phase 
instead of single phase current. This is done in his 1 5-ton 
furnace. One of these is operating at the works of the 
Illinois Steel Co., South Chicago, and another at the American 
Steel & Wire Co., Worcester, Mass. The hearth of this furnace 
is circular, over which the three electrodes are arranged at the 
corners of an equilateral triangle. The furnace is operated by 
three phase, 25 cycle current, delta connection, at 100 volts. 
Under these conditions, the current, per phase, rises to 12,000 
amperes. As in other furnaces the electrodes are automatically 
regulated. The current is taken from a high tension circuit and 
stepped down by means of three 750 kw transformers to the 
desired voltage of 100. Accordingly the furnace for 15 tons 

12000 X 100 X 1.73 = 2076 kva, and as the power factor is 
between .8 and .9, it consumes actually 

2076 X .85 = 1760, say 1800 kw. 

The difficulty of building one of these large arc furnaces lies in 
the increasing difficulty of finding a suitable electrode design, 
which will be durable in service and not have too great electrical 
or thermal losses, at the same time not injuring the general view 
of the hearth too much. This feature will be alluded to later on. 
In order to give an idea of the dimensions of the electrodes in 
Heroult furnaces, it may be well to mention that the electrodes 
carrying 4000 amperes in the single phase furnace operating at 
La Praz have a cross-section of 360 X 360 mm. = 129,600 sq. 
mm. (14.1 X 14.1 inches = 200 sq. in.), and a length of 1.70 

1 In 1912, the Metallurgical and Chemical Engineering reports that a 25- 
ton Heroult furnace was put in operation at the Gewerkschaft "Deutscher 
Kaiser," Bruckhausen, Germany. 


metres (67 inches). They consequently operate at a current 
density of 

129,600 / 200 

"4000" = 32 ' 4 Sq< mm * per amp * \4^> = - 5 Sq * in ' per amp ' 

4000 \ 

or == 20 amps, per sq. m.J 

If we take into account that as the height of the furnace roof over 
the bath is 70 cm. (27^ inches), and the clamping length at the 
top of the electrode is 35 cm. (i 3^ inches), we find that there is 
a certain length of usable electrode, which with a total length of 
1.75 metres (69 inches), makes the usable portion about 70 cm. 
(27^ inches). The w/msable portion of the electrode is con- 
sequently about i metre (39 inches). 

If we now calculate the electrode voltage losses in accordance 
with the figures just mentioned, similar to the electrode losses 
determined for the Stassano furnace, we obtain the following: 

Assume specific resistance of carbon in operative condition = 
Pi = .00183 ohms per cubic centimetre 

then as e = i X r, where r = pr~ , and / and q are respectively 

in centimetres and square centimetres, we obtain 

/ 100 
e == i X PI X = 4000 X .00183 X T = .565 volts. 

For both electrodes, then the drop is 1.13 volts, because they 
are connected in series. 

The result as figured, however, cannot be considered as correct, 
because the change in the specific resistance with increasing 
cross-section was not taken into consideration. In the calcula- 
tions, so far, we kept the probably correct value of .00183 ohm 
per cubic centimetre, which is in keeping for an electrode of 80 
mm. diameter, whereas the electrode of the Heroult furnace in 
question corresponds to a square having 360 mm. to a side. If 
this had been taken into consideration, then the value of pi = .014, 
(when following the values given on pages 93-94), should have 
been chosen for the electrode condition in its cold state. Should, 
on the other hand, the values of Hansen be taken, where the 


resistance falls to about 40% in operation, compared to the cold 
resistance, then the determination should have been figured with 
pi = .0056 ohm per cubic centimetre. 

Figuring more correctly then with this value, we obtain, 

e = 4000 X .0056 X T~ = i-73 volts per electrode. 

The drop for both electrodes is consequently 

2 X 1.73 = 3-5 VoltS. 

This gives a loss three times as high as in the first calculation. 
This example clearly shows of what importance it is to accurately 
know the different constants for this material for the different 
cross-sections. For it is only with these that the determinations 
of the conditions arising in the electrodes can be figured. 

Of course it is not to be supposed that this last value gives 
a final idea of the total losses in the electrodes, because in the 
calculations just made only the purely electrical losses were 
judged. This, too, with the rather hazardous assumption that 
the constant taken for the specific resistance of the carbon elec- 
trode is correct. Meanwhile, the radiation heat losses have been 
entirely disregarded. It is undoubted, that the latter raises 
the total electrode losses considerably, and even though deter- 
minations regarding radiation heat losses are hardly possible, 
still it may be said with some certainty, from measurements of 
other arc furnaces, that the total electrode losses generally, as 
well as in the Heroult furnace under discussion, will not be 
below 7 to 10%. 

These losses do not only appear of this value in the compara- 
tively small furnaces, such as have just been discussed, i.e., of 
the 2- to 3-ton size, but especially in the larger sizes. With the 
size of the furnaces and the increasing cross-sections of the 
electrodes, the difficulty also grows of obtaining favorable material 
constants, which is a thing entirely apart from the difficulties 
to be surmounted of procuring large electrodes of considerable 
durability. Chapter VI brings out these details. It may be 
said further regarding the practical operation of the 1 5-ton, 3- 
phase furnace at South Chicago and Worcester, that it has not. 


been found possible to increase the proportions of the electrodes 
at will. As has been remarked this 1 5-ton furnace operates 
with about 12000 amperes per electrode. The conduction of 
such currents naturally necessitates very considerable electrode 
cross-sections. It was at first tried to produce these electrodes 
in single large blocks. According to the Electrochemical and 
Metallurgical Engineering, 1909, p. 262, one of these block 
electrodes had a diameter of 2 ft. (60.9 cm.), by a length of ten 
ft. (3.048 m.). The weight of one of these electrodes was about 
32oolbs. (1451.5 Ag.). 

The results with these colossal electrodes was hardly satis- 
factory, as breaks often occurred which disturbed the operation 
of the charge in a most sensitive way, even though the current 
density operated with was 28 amperes per square inch, or 4.35 
amperes per square centimetre, corresponding to 24 sq. mm. per 
amp., which is a comparatively high density in spite of the large 
electrode cross-section. (See Chap. VI, page 82.) On that ac- 
count they sometimes use the dearer but less troublesome graph- 
ite electrodes instead of the carbon electrodes. Quoting from 
the Metallurgical and Chemical Engineering, 1910, p. 179, and 
following pages, we find that the electrodes as used are made up 
of Acheson graphite rods, 48 in. long (122 cm.) , and 8 in. (20.3 2 cm.) 
in diameter. Three such rods are butt-connected to a total length 
of 144 in. (366 cm.), and three such 144 in. rods are arranged 
side by side to form a single electrode, consisting (see Fig. 60 a 
and b) thus of a solid bundle of three rods, each 144 in. (366 cm.) 
long. The cross-section is therefore 3 X 50.2 = 150.7 sq. in. (3 X 
324 = 972 sq. cm.). The consumption of these electrodes is given 
as averaging 6.6 Ib. (3 kg.), per ton of steel, and this figure is 
stated to be true both for graphite and for amorphous carbon. 

The unavoidable wearing away of the comparatively dear 
electrodes, naturally causes an increase in the steel conversion 
costs, which is hardly desired. In the beginning there were 
additional losses of considerable moment which had to be reck- 
oned with. These were caused by the unusual lengths of the 
electrodes in the electrode clamps and the length necessary for 
the distance between the furnace roof and- the slag layer. These 


costs are said to have now been reduced to the irreducible 
minimum, by using the otherwise worthless stub ends for a new 
electrode. Figs. 58 and 59 show two possible ways in which the 
greatest use can be made of the electrodes. 

Fig. 58 shows the electrode made from shorter pieces with 
staggered ends held together with graphite screws. This method 
is also reported to have been used with the 1 5-ton furnaces in the 
United States. Fig. 59 shows a threaded hole in the end of the 
electrode. On the one hand this scheme enables the conducting 

clamp to be made of cast copper, as 
the figure shows, whereas otherwise, 
should the whole electrode become 
too short, it can be unfastened 
at the copper casting, a graphite 
screw inserted in its place, and a 
new electrode piece screwed be- 
tween the too short electrode and 
the new one. This is also evident 
from a view at Fig. 59. The latter 
way of lengthening the electrode is 
used at the Steel Works of Richard 
Lindenberg, at Remscheid, Ger- 
many. It seems from this that the 
possible difficulties due to the higher 

FIG. 58. FIG. 59. 

resistance at the points of contact are not so great as might be 
expected from theoretical calculations. 

Following this it may be well to relate further details of 
the operation of the Heroult furnace. If the Heroult furnace 
is to be heated up after putting in a new lining, or owing to the 
operation being interrupted by Sunday, it is accomplished by 
charging the furnace with some coke, which acts as the heating 
medium and at the same time as the conductor from one electrode 
to the other, (as long as the heating of the furnace is accomplished 
electrically,) before the charging of the regular metal. After 
the furnace is charged the electrode regulation is performed 
manually, especially when scrap is to be melted down. This, 
however, often occurs when hot metal is first charged when 


beginning operations. For instance, the 1 5-ton furnaces above 
mentioned are so regulated, whereas during the normal operation 
the regulation is also here accomplished by means of the Thury 

As with all other electric furnaces, so also with the Heroult 
furnace, we find that the power consumption varies greatly with 
the size of the furnace, with the kind of charge used, and the 
desired quality of the finished material. A graphic picture of the 
change of the current consumption varying with the size of the 
furnace is given by Fig. 60. This data is given by Eichhoff. 
Here one set of curves represents the conditions for cold and one 
for hot charges. In the upper set of three curves the lowest 
one indicates conditions when only one slag is used, the middle 
curve when two slags are used, and the highest curve when three 
slags are used, and similarly for the lower set of three curves. 
In this way the curves show a rising degree of purity in the 
metal. The table accompanying Fig. 60 gives the quantities 
directly. Of particular interest are the operating figures which 
have been achieved with the 1 5-ton furnace. According to the 
report in the Metallurgical and Chemical Engineering, of 1910, 
p. 179, ff., the electric furnace is charged with hot metal from 
the Bessemer converter. 

On the average here 12 charges are made daily, with an average 
time of i hr. and 15 minutes to 2 hrs. and 15 minutes, the weight 
of metal averaging from 12 to 14 tons. The average consump- 
tion of power for this is 200 KW. per ton of steel produced. 

If we now pass on to the comparison between the Heroult 
furnace and the ideal furnace, we come to the first demand of an 
electric furnace, that every existing alternating current can be used. 
The Heroult furnace fulfils this demand only in part. As every 
arc furnace needs a certain voltage, the Heroult furnace also 
demands a specific potential, so that in nearly every instance a 
stationary transformer becomes a necessity, in which the high 
pressure of the distant central station is stepped down to the 
desired 100 to no volts at the furnace. The use of one of these 
transformers is almost unavoidable with any arc furnace. We 
saw further that up to the present the 1 5-ton Heroult furnace 



is operated with three phase current, whereas all his smaller 
furnaces are operated with single phase current. As it has proven 
possible to operate large furnaces with three phase current, 
it is obvious that the building of three-phase Heroult furnaces 
is constructively feasible. If, in spite of this, the building of 
such three phase furnaces has not taken place, the reason can 
only be sought in operating difficulties, be they electrical or 
metallurgical. Furthermore, with a new power plant nobody 
would decide to install the dearer single phase machines, if one 
could operate successfully with the cheaper three phase machines 
which are often already installed. It is not very hard to find the 
causes which have made it seem non-advantageous to use three 
phase current for normal Heroult furnaces of 2 to 5 tons capacity. 
The arrangement of three electrodes with their regulating 
mechanisms on furnaces of this size, makes the whole outside 
appearance of the furnace less simple and obstructs the general 
view, which is emphasized even more when considering the 
obstructed view of the comparatively small hearth. On the 
bath of these hearths three electrodes would appear which 
would hinder the metallurgical operations and lead to greater 
breakages of the electrodes, than two electrodes alone would, of 
which it is reported that breakages are rare during operating. 
In addition the arrangement of three electrodes would require 
the furnace roof to be pierced three times, which seems so much 
more dubious, as the arched roof is subject to the high tempera- 
tures of the arcs, and also to the water cooling around the three 
electrodes, so that inside of the comparatively small space of the 
furnace roof, we would have several large differences of tempera- 
ture arising, which naturally tend to weaken and destroy the 
roof. Finally it must also be mentioned that three electrodes 
radiate more heat due to their larger surface than two electrodes 
do, having the same total cross-section. 

These reasons show us, therefore, why 5-ton Heroult furnaces 
at present are only operated by single phase current. The 
consequence of this being that, when furnaces up to 5-ton 
capacity are to be installed where three phase current is available, 
a three phase single phase motor generator becomes necessary. 


Whereas, if a new central station is to be built, it would be cor- 
respondingly dearer, as single phase generators are costlier than 
three phase generators. Larger furnaces which can be operated 
with three phase current consequently only require the installa- 
tion of stationary transformers, provided the central station is 
large enough to stand the prevailing power fluctuations. 

Up to the present time the avoidance of sudden power fluctua- 
tions has not been attained when operating Heroult furnaces. At 
present the Heroult furnace, with its electrodes in series, is 
credited with having the heaviest power fluctuations of the 
better known arc furnaces. This, as we have seen, is particularly 
so when melting down cold stock, during which time the fluctua- 
tions often become so great that the automatic regulation fails, 
and the hand regulation has to be resorted to. The conditions 
are more favorable as soon as the charge is completely melted, 
or when only treating hot charges. Once again, reference may 
be made to the curve shown by Fig. 57. 

Easy regulation of the power is present in the Heroult furnace, 
the same as it is in every other electric furnace. 

In judging the electrical efficiency of the furnace, the losses in 
the transformer are first to be taken into consideration, and 
then the losses in the carbon electrodes. In case any rotary 
transformers have to be used, the considerable losses appearing 
here have to be added. 

In order to give a probable conception of the electrical losses 
the efficiency of the transformer may be taken as about 96 to 
97%; the electrode losses at about 10%, of which at least 3 to 
5% are purely heat losses, and in case rotating transformers have 
to be used the efficiency of these machines may be taken at about 


The further requirements of a tilting furnace, and an easily 
surveyed and accessible hearth are fully met. 

It has already been mentioned in which way Heroult knew 
how to avoid the undesired reducing action of the electrodes 
:"mpinging directly on the metal. It is to be noted, however, 
that this reducing action cannot be altogether avoided, due to 
the electrodes throwing their carbon vapor stream against the 



layer of slag, even though the slag layer protects the bath from 
this action. The prolonged carbonizing action of the arc furnace 
makes more difficult the oxidizing processes; for instance, during 
the removal of the phosphorus it cannot be without its influence 
on the time of treating the charge and the power consumption. 
When removing the slag it is well to consider that the carbonizing 

FIG. 6oa. Heroult 3-phase furnace of 1 5 tons capacity. Teeming a charge. 

action of the arc remains the same, even though the heating is 
not interrupted during this period. 

If we now take up the requirement of the motion of the charge, 
we find that from reasoning alone, from the standpoint of purely 
thermal action, it is not present. For, as the arc operates only 
on the surface of the bath, the hottest parts of the bath are to be 
looked for here. On account of the electric current, on the other 
hand, a certain motion of the bath takes place, as this current 


flows through the electrodes, and a part of the bath which is, 
to a certain extent, a moving conductor (as the motor action of 
the electric currents acts as discussed in Chapter III). For, in 
accordance with the conditions there given, the bath, or the part 
which is a movable conductor, is pushed to one side, so that the 
material beneath the electrode is under certain magnetic 
pressure, which causes a certain motion in the bath of the Heroult 
furnace. With all this, it is not correct to assume that the 
motion caused in the bath of the Heroult furnaces reaches the 
bottom of the bath. 

The application of the furnace has a wide scope, although it 
is restricted from the melting of cold stock, due to too great 
fluctuations of the power, or if in other cases the installation of 
a rotary transformer is not considered prohibitive. However, it 
must not be left unsaid that at present the Heroult furnace is 
the only one which has been built for a charge of 15 tons, which 
proves the adaptability of the furnace for this size. Up to a 
certain point naturally the heat losses become proportionately 
smaller. However, it is to be feared that, for instance, with 
very large arc furnaces with three electrodes the furnace 
roof will have to be renewed quite often. The renewal of these 
new roofs for the usual 3 to 5 ton size furnaces is not exactly 
pleasant, to say nothing of the cost entailed. These renewals 
are, however, to be feared even more with the large furnaces 
with three openings in the roof, for besides the long span of the 
roof, there are the large differences of temperature between the 
various parts of the covering. This disadvantageous trait 
remains even though it is considered that the vertical electrodes 
act with a sort of umbrella action, and in so doing at least keep 
the most intense heat away from the roof. It is to be noted 
here that, for instance, the roof of the 1 5-ton furnace at the 
Illinois Steel Co., has to be changed every Sunday. According 
to the Metallurgical and Chemical Engineering, a roof such as 
this costs about $60. If a longer life is desired of the roofs of 
the Heroult furnaces, then it seems that the only course left is 
to make the covering as high as possible above the hearth. This 
would, however, cause larger electrical losses, as the current 



would have to travel along a greater length of electrode, which 
would also cause greater heat losses, as these grow with the 
enlargement of the free space between the bath and its arched 

All these reflections seem to point to the 1 5-ton size as the 
largest Heroult furnace to be recommended at present. This 

FIG. 6ob. Heroult 3-phase furnace of 15 tons capacity, at Worcester, Mass. 

may be emphasized even more, consider 'ng the serious difficulties 
which have been encountered with the electrodes obtainable 
today for this size furnace. That these difficulties have not yet 
been overcome may be judged from the report appearing in the 
Metallurgical and Chemical Engineering for 1910, p. 1796"., 


where the electrode temperature, just where it issues from the 
furnace was measured and gave 1050 C. It is evident that 
these electrode temperatures cause a greater consumption of 
the electrodes, so that this may also be looked upon as part of 
the cause for the high consumption of 6.6 Ib. (3 Kg.) of electrodes 
per ton of steel. 

It is also to be noted that it must be possible to change the 
slag in an electric furnace, as is now done in the open hearth 
furnaces. The removal of this slag, however, becomes more 
difficult with the increase in the size of the furnaces, because the 
slag must be entirely removed. A mere running off of the slag 
is not sufficient, but a thorough rabbling off is necessary. In 
taking these conditions into consideration the Electrochemical 
and Metallurgical Industry of 1909, p. 262, says in referring 
to the attainable size of the Heroult furnace: 

"As to the maximum size of furnace which it is now possible 
to construct, it is the intention to build them up to 30 tons. 
Very much will depend however on the work which has to be 
accomplished, that is to say, whether one or two slags would 
be used. In case of one slag, Mr. Turnbull is sure that a 30-ton 
furnace is possible, but should two slags be used, owing to the 
difficulties which might be encountered in raking off the first 
slag it may be found that a 1 5-ton capacity is nearing the limit. 
It could certainly be worked quicker than one of a 3o-ton capa- 

Attention is again called here to the influence of the furnace 
size on the thermal efficiency of Heroult furnaces, and this point 
is dwelt upon more in detail. Prof. Eichhoff says the following 
in Stahl und Eisen, 1908, p. 844: 

"I cannot think of a small furnace that has an efficiency of 
more than 50%. If the furnaces become larger and larger, 
then the actual useful absorption of the heat may rise to 70%, 
for the reason that the furnace surface does not increase in the 
same ratio as the furnace contents do. As the furnaces become 
larger the losses gradually decrease going from 50 to 40, and from 
30 to 25%. I can tell you from my own practical experience, 
that comparing a 3-ton furnace to a i. 5-ton furnace, the effective 
current increase was only 10%. Hence, the current consump- 


tion per ton of steel decreases materially. Owing to this fact 
we are compelled to build larger furnaces, and there is no reason 
why this cannot be done." 

Since then there has been built the furnace of 15 tons, 
as mentioned by Eichhoff. For this size the above deductions 
are correct, however, with the limitations that the furnace 
efficiency cannot be further increased by further increasing 
the size of the furnace unit. The efficiency of furnaces of 
increasing sizes with two electrodes follows the curve of a 
parabola. However, where three electrodes are used, the 
efficiency will naturally decrease, due to the higher thermal losses, 
which latter gradually reach the practical attainable minimum, 
with the increasing size of furnaces. As Heroult furnaces, 
however, are built today, these losses will not be less than 25%. 

It is difficult to calculate definite Heroult furnace installation 
costs, as these will in all cases be determined largely by 
local conditions. The direct connection of a single phase 
Heroult furnace, to an existing power plant, will hardly ever 
occur as the latter are usually of three phase design. Hence, the 
single phase furnaces necessitate a single phase, three phase 
rotary phase changer which is more expensive than the furnace 
itself. Comparing Heroult furnace installations with others, 
the former appear at their best when each installation has its 
own generator set. Besides this, as has already been pointed 
out, it is necessary to install a low tension regulating transformer 
near the furnace, which lowers the higher potential of the central 
power plant. 

In giving the installation costs of some Heroult furnaces, we 
may figure roughly with the following values: 

For a 5-ton Heroult furnace, a special single phase generator 
at 25 cycles of 600 KW. would be necessary, costing about $9,000, 
including foundations, erection and controlling switchboard. 
This, however, does not include the prime mover. To the above 
cost comes the furnace transformer, the transformer apparatus 
for regulating the electrodes, the cables and fittings, and the 
furnace itself with its regulating and tilting mechanisms. This 
will entail an additional cost of $12,000. 


We would, therefore, have to reckon with $21,000 for a 
5-ton Heroult furnace installation, including the generator, 
provided it is not desired to connect to existing three phase power 
mains. If, however, this becomes desirable for any reason, then 
a motor generator must needs be added, increasing the above 
prices by $6,000, so that in this case the total cost would be 
about $27,000. 

In all the above calculations, the cost of buildings, etc., are 

In closing, the advantages which Heroult himself gives of 
his furnace, over other arc furnaces, are here set down, especially 
those opposed to the Girod furnace, which latter is described in 
the following chapter. The advantages mentioned are taken 
from the Electrochemical and Metallurgical Industry, for 1909, 
p. 261: 

" First The total absence of electrical parts in the furnace 
proper, it being nothing else but a modified open hearth with 
the heat introduced above the metal by the electric current in 
place of gas. This in itself is an important factor as it does 
away with the bottom pole, considered by Heroult to be always 
the cause of much trouble in electric furnace work, and allows 
of any patching necessary to the bottom or side, without inter- 
fering with the work of the furnace. 

" Second The heat being introduced by means of two 
electrodes working in series, the current passing through the 
bath from one electrode to another and vice versa, necessitates 
carrying only one-half the current that would be the case should 
the current flow from one electrode through the bath and then 
through the bottom of the furnace, if the power is the same in 
both cases. Thus, all the conductors are reduced to one-half the 
section required in the other case and the electrodes can perform 
more efficient work owing to the lesser density of current to be 

The above-mentioned advantages of the Heroult furnace 
should be compared to the advantages of the Girod furnace, 
mentioned at the end of the following chapter. Furthermore, 
the opinion in the first paragraph may be supported. It is 
correct, of course, that certain advantages accrue by lessening 


the cross-section of the current carrying conductors. He, how- 
ever, avoids mentioning that these advantages are only attainable 
by raising the voltage. Even though a no- volt alternating 
current pressure is usually harmless, yet sensitive persons are 
apt to experience most uncomfortable shocks. For this reason 
it is commendatory that the operators of a Heroult furnace 
protect themselves against these shocks, by, say, wearing wooden 
soled shoes which are insulators when dry. The opinion of 
Heroult that the series connection of the electrodes gives more 
useful work, is not substantiated in any way. We shall see later 
on that the total electrode cross-section of the Heroult furnace 
is not greater than with the Girod furnace, disregarding entirely 
how incomprehensible it is that Girod does not also operate 
with the same current density and the same low current densities 
as Heroult does. It still remains to be proved that operation 
with low current densities is an advantage, irrespective of the 
size of the furnace. 

Relative to the use to which the Heroult furnace has been 
put, reference may be had to the statistics in the closing chapter. 
Licenses for Heroult furnaces may be obtained in Germany from 
the Elektrostahl, G. M. b. H., Remscheid, Hasten, and in the 
United States from the United States Steel Corporation, New 



THE Girod furnace, as well as the Heroult furnace, deserves 
the greatest consideration among arc furnaces. Girod originally 
made ferro alloys in a resistance furnace, in which the heat flow 
went through the walls, as described in Chapter III. It was in 1906 
and 1907 that he turned quite experimentally to the melting of 
iron. He built a furnace with a capacity of about i to i^ tons of 
a similar type to that used by Heroult, before the latter went over 
to his electric furnace with series connected electrodes. Where 

FIG. 61. 

Heroult did not succeed in obtaining satisfactory results with his 
furnace, having one pole in the form of a hanging electrode, and 
the other pole as a bottom electrode, Girod succeeded. Girod 's 
success has been so great in bringing this furnace to such a fully 
developed scientific reality, that it is hard to say at present to 
which of these two contestants, in the arc furnace field, the 

victory will finally belong. 




In outward appearances the Girod furnace greatly resembles 
the Heroult furnace. The furnace casing is made of steel plate 
and either of the round or rectangular form. This in turn re- 
ceives a lining of either dolomite or magnesite, making the bath 
either round or square shaped, as the case may be. The furnace 
roof is made of silica brick and is removable. The furnace itself 
is of the tilting variety. Because of this the first furnace at 
Ugine, France, was provided with trunnions at the side, which 
allowed the furnace to tilt in its bearings. In the newer design 
the furnace casing is furn'shed with a saddle resting in rollers, 

FIG. 62. 

as shown in Figs. 61 and 62. The power for the tilting mechan- 
ism may be of any kind, but is usually an electric -motor. The 
Girod furnaces are supplied with two doors, one of which serves 
mainly for the charging and operating of the furnace, while the 
other is provided with a teeming spout, for the tapping of the 

The most interesting part of the Girod furnace is, of course, 
the arrangement of the electrodes in which centres the whole 
principle of the furnace. Where in the Heroult furnace the 


electrodes are of opposite polarity and arranged above the bath, 
Girod avoids this by placing one pole above and one beneath 
the bath. When the current strength increases with larger 
furnaces, and a duplication of the electrodes becomes neces- 
sary, then these are connected in parallel. This always per- 
mits electrodes of the same size to be used, and like poles are 
therefore either only above or below the molten metal. This 
arrangement, which naturally only allows the electrode above 
the bath to be of carbon, from which the current flows to 
the liquid steel in the form of an arc, allows the other pole lying 
beneath the bath to be of a special formation. In the Girod 
furnace this bottom electrode consists of a number of soft iron 
rods, which are arranged at the edges of the hearth, as seen in 
the horizontal cross-section of Figs. 61 and 62. In order to avoid 
these bottom electrodes from melting off too far, the parts pro- 
truding through the furnace bottom are water cooled. During 
the operation then a part of these electrodes melts away, after 
which pasty layers, followed by solid ones, issue toward the 
bottom of the electrode material, as soon as the cooling on one 
side is balanced by the heating on the other. The part of the 
electrode which is melted away is about 5 to 10 cm. (2 to 4 inches) 
long, whereas the space for the water-cooling at the lower end 
of the iron block is 150 mm. (6 inches) deep. This water cooling 
not only provides a nearly unlimited durability to the bottom 
electrodes, but it also materially aids the life of the bottom 
refractories. From data given by Borchers, the furnace bottom 
is said to last 120 to 160 heats when melting cold stock, before 
repairs are necessary. During this time the bottom wears away 
to the extent of 100 mm. (4 inches), whereas the walls of the 
furnace need repairing after only 80 heats. 

It may also be mentioned here, that Girod endeavored to 
utilize air cooling in place of water cooling for the bottom elec- 
trode, but at present water cooling is again generally used. 

What has been said of the Heroult furnace relative to the 
hanging carbon electrode also applies here. The adjustable 
electrodes are held in their supports, which are in turn fastened 
to the furnace. The regulation is automatic and the Thury 


regulators are used. Another similarity is to be found in the 
method pursued for cooling the furnace roof, where the electrodes 
enter the furnace. 

The operation of the furnace and, with it, the duration of 
the treatment, is much the same with the Girod furnace as with 
the Heroult. This applies as long as hot charges are being 
treated, for when it comes to melting cold charges, the Girod 
furnace shows undeniable advantages over the Heroult furnace. 
This is because the vertical path of the current does not permit 
any short circuits at almost full voltage, when the upper electrode 
touches the top of the scrap pile. When the electrode is lifted 
clear of the furnace, the scrap entirely fills its interior, and the 
short circuits are avoided, as the current path necessarily makes 
a multitude of small arcs between the various pieces of scrap. 
This equalizes the heating of the whole furnace content, thus 
causing the whole charge of scrap to gradually collapse and melt. 
However, it must not be left unsaid that the above conditions 
are present only when the scrap is charged into the furnace as 
the best operating conditions of the furnace demand; that is, 
the scrap is not to be thrown in arbitrarily. The most advan- 
tageous condition for melting cold stock is when this is in the 
smallest of pieces, and the conditions become more disadvan- 
tageous with the growing number of larger pieces. For these 
latter offer far too little resistance to the current, if the above 
method were used by starting with the upper electrode touching 
the top of the scrap pile. Similarly it is always necessary to 
spread a layer of the smallest sized scrap on the hearth, so that 
good contact can be made from the start with the bottom elec- 
trode, the end of which naturally lies a little low after the furnace 
has been in operation for a while. In order to make a good 
contact possible between the bottom electrode and the charge, 
care must be taken that no slag remains in the indentation over 
the iron electrode, otherwise this cold slag would act as a con- 
ductor of the second class, and in this state act as an insulator. 

We now come to the electrical conditions of the Girod furnace. 
Heretofore this furnace has been built mostly in two sizes. The 
smaller size of 2^ tons capacity shown by Fig. 61 and the larger 


size of 10 and 12 tons shown by Fig. 62. The smaller furnace 
takes about 300 Kw. and the larger from 1000 to 1200 Kw. As 
the current is only interrupted by one arc the resistance of the 
whole circuit of the Girod furnace is comparatively small. From 
this it follows that a comparatively low voltage suffices, in order 
to give the furnace its needed energy. The voltage therefore 
for the 300 Kw. furnace is from 60 to 65 volts, and with the 1000 
to 1200 Kw. furnace it is 70 to 75 volts. 

As the furnace has only two poles, one above and one below 
the bath, naturally only single phase current can be used. As 
the Heroult furnaces are operated almost exclusively from 25 
cycle circuits, so the Girod furnaces today operate exclusively 
from circuits of this periodicity. The first trial furnace of i^ 
tons tapping weight operated from a 35 cycle circuit, using 40 
to 60 volts, 4000 to 6000 amperes and giving a power factor of 


The low voltage of the Girod furnace naturally necessitates 
a comparatively large current, and with it very considerable 
cross-sections in the conductors between the furnace transformer 
and the furnace. This is very noticeable when comparing the 
furnace with a Heroult furnace having an equal charging capacity 
and the same power input. It is this lower voltage which makes 
this part of the installation more expensive than would be the 
case with a furnace having a higher operating voltage. We 
must, however, take into consideration that the lower voltage 
also has its advantages. We only mention the fact that it is 
easier to insulate this voltage from the furnace refrac- 
tories, and there is less danger for those operating the 

It has already been remarked that the Girod furnace is the 
youngest among the better known arc furnaces. It is therefore 
not to be wondered at, that it is not yet absolutely clear how best 
to operate the furnace. This is why we find in the comparatively 
sparse literature on the Girod furnace, the recurrent opinion that 
the Girod furnace is radically different from the Heroult, owing 
to the fact that the bath is connected in the circuit in a different 
way. We have already alluded to the advantage of the current 


passing through the steel and iron in a vertical direction, when 
melting cold scrap. We desire, however, to discuss the operation 
of the furnace when the charge is melted. 

As we have seen, the current passes through the bath in 
a horizontal direction, in the Heroult furnace, and in a vertical 
direction in the Girod furnace. With the Heroult furnace, 
however, mention is never made of any essential influence of the 
purely resistance heating, which occurs because the current must 
overcome the resistance of the bath, yet with the Girod furnace 
we often find an important heating effect ascribed to it. In order 
that there shall be no misunderstanding, it may be said that the 
different manner in which the current goes through the bath in 
both furnaces causes different effects, yet these effects do not 
cause a greater or less resistance heating, (caused by the current 
passing through the bath,) but rather a difference in the circula- 
tion phenomenon. This may be decidedly more advantageous 
in one case than in another. To which misleading points of 
view our opinions lead us to suppose that the resistance heating, 
(even with a molten bath,) is of considerable influence, is 
shown in a short article on the Girod furnace in Stahl und 
Risen, for 1908. Here it is pointed out that the depth of the 
melted iron of the Girod furnace may easily be increased from 
30 cm. (12 inches), to 75 cm. (30 inches), or more. With all this 
the pure resistance heating is supposed to heat the whole bath 
evenly throughout its total depth. In spite of this, though, 
Girod with his 10- to 1 2-ton furnaces only used a depth of bath 
equal to 30 cm. (12 inches). 

A large surface bath has much greater radiation losses as a 
consequence than a bath has, having great .depth and a lesser 
surface. The above example of the 12 -ton furnace really proves 
that the resistance heating in a Girod furnace can be entirely 
ignored, as soon as the furnace content is molten. We can also 
convince ourselves of this arithmetically. 

If we take, for example, the 2^4-ton. Girod furnace, we find 
by consulting Fig. 61, that with a depth of bath equal to 240 mm. 
(9.1 inches), the average bath cross-section is about 1200 X 1200 
sq. mm. (48 X 48 sq. inches). If we take the specific resistance 


of the bath at 1.66, as given on page 15, we find the ohmic resist- 
ance of the bath, 

= p X -- = 1.66 X - - = .28 X 10-6 ohm. 

q 1200 X 1200 

A furnace of this kind takes about 300 Kw. at 60 volts. 
With a power factor of .8% it gives a current of 

6o~^T = 625 amperes ' 

The energy, therefore, transformed in the bath is: 
i~ X r = 625o 2 X .28 X io~ 6 = 10.94 watts. 

This amount is only - of i% of the 300,000 watts delivered 

to the furnace, and everybody must admit that any such small 
amount of energy has absolutely no effect on the heating. If, 
on the other hand, we figure the current density in the bath, we 
will see that this comparison also shows the heating of the molten 
metal to be entirely uninfluenced by the current flowing through 
this resistance, and that the resistance heating of the carbon 
electrodes is much more important than the resistance heating in 

the bath. The example we have before us gives, - - = 

' 1200 X 1200 
6250 ( 6250 

- = .0044 amperes per square millimetre s 
1,440,000 ^48 X 48 

6250 2.71 amperes per ) 

= . . > which allots 230 sq. millimetres to 

2304 square inch ) 

i ampere (about .36 sq. in. per ampere). If we compare this 
with the current density in the carbon electrode, which con- 
ducts the same current that flows through the bath, and has 
a cross-section corresponding to a diameter of 350 mm. or 
96211 sq. mm. (13^ inches dia. gives 149 sq. inches) with the 
300 Kw. furnace, we observe that we only obtain a cross-section 

.96211 / 14-9 .024 square in. \ 

per ampere of -- = 15.4 sq. mm. I- -- = ) 

6250 ^6250 per ampere. / 

With all this it is well to note that the comparison of 
these absolute values gives a much too favorable picture, because 
no consideration has been taken of the higher specific resistance 
of the carbon compared to the iron bath. In accordance with 


data on page 15, we figured the specific resistance of fluid iron 
as p = 1.66. This resistance refers to a length of i m. (39.37 
inches), and i sq. mm. (.0155 sq. inch), cross-section. With 
electrodes in the operating condition we figured p = .0056 (see 
page 130). This value corresponds to a length of i cm. (.4 inch), 
at a cross-section of i sq. cm. (.155 sq. inches). If we convert 
this value to one corresponding to a length of i m. (39.37 inches), 
with a cross-section of i sq. mm. (.0155 sq. inches), we acquire 
the value for carbon in the operating condition, when p 56. 

That is to say, the specific resistance of the carbon is 

i .00 

or say, 35 times larger than that for iron. From this it follows 
that even with equal lengths and cross-sections, 35 times as much 
energy is transformed into heat in the carbon as in the iron bath. 
It is obvious that the carbon has a much smaller cross-section 
and a much greater length than the metal has, evincing that the 
consequent heat distribution is much more unfavorable for the 
iron, when considering only the resistance heating. The true 
ratio is therefore not apparent by the above partial calculation. 

We will now consider the comparison of this furnace with 
the ideal furnace. We first come to the availability of any kind 
of alternating current and refer again to the former remarks, that 
single phase current is only available for this type of furnace, as 
the use of three phase current would come in conflict w'th the 
principle underlying the furnace design. In case a single phase 
generator is not specially installed in the power-house, and 
should it be desired to operate from an already existing power 
system, then it is necessary to furnish a rotary transformer. 
This latter would then convert the prevalent three phase current 
or direct current into the desired single phase alternating current. 
Alternating current is usually generated at a commensurately 
high voltage, brought to the vicinity of the furnace, and there 
transformed :'nto a stationary transformer to the wished-for low 
tens 'on current for the furnace. 

It is just as difficult to entirely avoid the power fluctuations 
with a Girod furnace as it is with a Heroult furnace ; yet it is to 
be observed that with the Girod furnace the current fluctuations 


in actual practise are neither as violent nor do they occur 
as often as they do in the Heroult furnace. In spite of the 
current fluctuations being smaller, they are yet important 
enough in a 2^-ton Gi od furnace which takes 400 Kw. on an 
average with a power factor of .80%, to recommend that a 500 
to 550 Kw. machine be employed. This example may properly 
show why the initial cost rises which really becomes noticeable 
here, all due to these power fluctuations. Finally, we may 
again mention, that the automatic regulation is accomplished 
by means of Thury regulators. With Girod furnaces, these 
regulators are set to keep the current constant, and they in turn 
give the electrodes their proper setting. 

The easy regulation of the incoming energy in tne Girod 
furnace is the same as with all other electric furnaces. 

The electrical efficiency of the furnace is influenced, first by 
the probable installation of a rotary transformer, latterly, by 
the losses of the stationary transformer, (neglecting the losses 
in the conductors,) and finally by the losses at the furnace, due 
to the electrodes. For a general calculation we can use the 
following values: 

Efficiency of the rotary transformer 85% 

Efficiency of the stationary transformer 96% to 97% 

Efficiency of the carbon electrodes including the heat 

conduction losses 90% 

All Girod furnaces are made of the tilting variety. The 
hearth is easily surveyed, and perfectly accessible for all operating 

We now come to the circulation of the melted metal and once 
more to the fact that the peculiar path of the current in a Girod 
furnace is of added importance. This circulation begins with 
one or more current centres above the bath, and goes to the 
bottom electrodes set around the periphery of the furnace. Figs. 
63 and 64 show the diagrammatic connections for the current 
paths in a Girod furnace, Fig. 63 being the plan view and Fig. 64 
showing the cross-section at a b. The dots and crosses indicate 
the lines of force, which follow the arrows according to the laws 
given in Chapter III. As lines of force of the same direction 



repel while those of opposite direction attract, and as the molten 
bath in a certain sense can be regarded as a movable conductor, 
with the vertical electrodes over and under the bath considered 
as fixed conductors, we find in the molten steel certain circulation 
phenomena, as shown by the arrows in Fig. 64. That is to say, 
a definite circulation will appear throughout the entire bath, of 
such a nature, that a current of metal can be observed going 

FIG. 63. 

FIG. 64. 

from the walls of the furnace toward the centre, from there to 
the bottom, and back again to the walls. The strength of this 
circulation phenomenon depends on one hand on the strength of 
the current which flows through the bath that is then collected 
at the electrodes, and, on the other hand, on the depth of the 
bath. For it is evident that the circulation in the bath would 
instantly cease if the metal currents were in a vertical direction, 
instead of being in an almost horizontal direction. If the bath 
has a comparatively great depth, we would approach the vertical 
direction condition. We ascertained before, that as the heating 
in a Girod furnace is practically entirely from the arc, a great 
depth of bath is therefore precluded, so we see now that the 
advantageous mixing in the bath would cease if this were, say, 
over 40 cm. deep (about 16 inches) this adequate mixing being 
present with shallow baths or those of normal depth. 

The application of the Girod furnace for the steel industry 
is one of the widest. It has already been said that very good 
results are obtained with furnaces of the 1 2-ton size. Here 
however, they already use four electrodes of considerable cross- 


There is then no reason why Girod furnaces cannot be built 
of the same capacities as, for instance, the Heroult furnaces, 
even though the Girod furnaces operate with a lower voltage than 
the Heroult furnaces, and although the latter operate with three 
phase current which has not yet been used for the former. In 
order to show this, we will assume having 1200 Kw. energy at 
.80% power factor, to be used by means of three phase current 
at no volts on the one hand and at single phase current at 70 
volts on the other, in the former case for a Heroult and the latter 
case for a Girod furnace. 

Then, per phase, we obtain for the three phase Heroult 
furnace, a current of 


^ h = ^, = 7882 amperes, 

" no X 1.73 X .8 

.and for the single phase Girod furnace the current: 

I 200000 

* = TolTTs = 2I429 am P eres - ' 

Now Heroult has to deal with 7882 amperes for each phase, 
i.e., three electrodes are needed each to cary 7882 amperes, 
whereas Girod has only to carry once a current of 21429 amperes. 
Suppose we assume that he too uses three electrodes, connected 
in parallel of course, then each would carry a current equal to 
21429 -T- 3 = 7143 amperes. In other words, it would even 
suffice Girod to have a lesser total electrode cross-section than 
Heroult, though the latter has a much higher current in the 
electrodes at the same current density. Or we may say: "The 
influence of carrying the current in, one way or another, is of so 
little importance as regards its effect on the carbon electrodes, 
and that the electrode relation in both types of furnaces may be 
regarded as being exactly like." Therefore, the same reasons 
govern'ng the maximum size of the Heroult furnace cover the 
Girod furnace also, so that the attainable size of either furnace 
is on the same footing. A certain limitation of the applicability 
of the Girod furnace may arise in certain cases, as the furnace so 
far has only been built for single-phase current, which necessi- 
tates expensive rotary transformer units for large furnaces, if the 


current of an existing three phase central station is to be used. 
If no consideration need be taken of an existing power plant, 
even then the single phase generators for large Girod furnaces 
will be more expensive than three phase machines of the same 
size for Heroult furnaces. Regarding the uninfluencing effect 
of the electric heating on the chemical composition of the bath the 
comment given on page 137 is also applicable here. This 
applies to all arc furnaces which have their electrodes directed 
directly against the metal to be treated. 

Especially worthy of mention with the Girod furnace is the 
influence which the water-cooled bottom electrode exercises, 
even though this influence is said to be of no consequence. To 
understand this, consider that the circular motion in the 
bath also continually renews the coldest material over the 
bottom electrode, so that in spite of the greater temperature 
difference between the bath surface and hearth bottom, 
there remain practically the same conditions as in the Heroult 
r urnace. 

In coming now to the consummate efficiency of the Girod 
furnace, it may be again said that, compared to the Heroult 
furnace, the proportions of the carbon electrodes in both 
furnaces may be looked upon as being equal to each other. From 
this it follows that not only are the electrical losses equally great, 
but the thermal losses also, for these are caused by the hanging 
carbon electrodes. Also, the water-cooling losses, caused by 
the devices at the roof of the furnace, where the carbon electrodes 
pierce it, are by no means unimportant. According to the report 
of Conssergues, these are about 10%; for it was established that 
the power consumption with the Girod furnace decreased 10% 
when it was operated without the water cooling. The fact that 
water cooling apparatus is used today on all Girod furnaces as 
well as on all other arc furnaces, may be explained by the follow- 
ing reasons, (as discussed on page 100,) first, a tighter fit can be 
made at the water cooling entrance to the furnace, the electrodes 
being better protected against oxidation, and, secondly, because 
the water-cooled boxes allow the furnace roofs to be stiffened, 
which latter have their life considerably prolonged. Besides 


these roof and wall radiation losses of the furnace which are 
about equal in the Girod and Heroult furnaces, there remain 
still to the detriment of the Girod furnace the losses of the water- 
cooled bottom electrode. These are avoided in the Heroult 
furnace. 1 We come to the conclusion, therefore, that the losses 
due to the cooling of the bottom electrode are according to 
an address by Trasensters "much less important," than those 
which are occasioned by the cooling where the roof is pierced 
for the carbon electrodes. 

In order to further judge the total efficiency of the Girod 
furnace, the following notation is taken from a report of the 
firm, Ohler & Co., of Aarau, in Switzerland. (See Electro- 
chemical and Metallurgical Industry, 1908, pp. 452 and 453.) 
Here we first find a description of Girod furnace installation at 
the above works. The furnace is connected to the power of 
the municipal power plant, through the medium of a motor- 
generator set. The 2000 volt, 2 phase current system supplies 
the Ohler Works' motor of 450 HP, running at 560 R.P.M., 
and is coupled directly to a single-phase alternator giving 4600 
to 5000 amperes at 65 to 75 volts and a frequency of 37.4 periods 
per second. Twelve heavy copper cables, each 20 mm. in 
diameter and composed of 12 copper wires twisted together, 
carry the current 10 metres to the furnace. The voltage drop 
is 2.5 volts from the machine to furnace, so that this short cable 
installation alone causes a loss of 3 to 4%. At the end of this 
report we find this statement. It is calculated that the elec- 
trical part of the plant has an efficiency of 75 to 80%; i.e., 75 
to 80% of the energy of the primary current appears as heat 
in the furnace. A rather approximate estimate of the calorific 

1 With the same construction of the Girod furnace as the Heroult, other 
things being equal, the efficiency of the Girod furnace must be just that amount 
less, which corresponds to the water cooling of the bottom electrode. Accord- 
ing to Stahl u. Eisen, July 20, 1911, by A. Miiller, in a 3-ton Girod furnace, a 
calorimetric determination of the heat carried out in the cooling water of these 
bottom electrodes gave 10.1 kilowatt-hours for the 130 minute run and about 
i. 01 per cent., or 2.9 kilowatt-hours per ton of steel produced. The cooling 
water used in the top electrode carried out 36.7 Kw. hrs., 3.65% of total 
energy supplied or 10.5 kilowatt-hours per ton of steel. 


efficiency of the furnace itself shows about 50% of the current 
converted into useful heat. 

Naturally the efficiency with the Girod furnaces also rises 
as the furnace increases in size. Yet, it is to be noted that, when 
the number of the upper electrodes is increased, the efficiency 
curve decreases here as well. The reasons underlying this were 
given in the preceding chapter on the Heroult furnace. 

The costs of a 2^-ton Girod furnace including the electrode 
regulators, the switchboard instruments, the tilting mechanism, 
its motor and short conductors between the furnace and its 
transformer or the dynamo room, total, according to Borchers, 
about $3,000. A large furnace of 10 to 12^ tons with the same 
equipment will cost about $7,000. 

The cost of a complete Girod furnace installation, but ex- 
clusive of the transformer or generator, and consisting of an 
operating and a reserve furnace each of 2 tons capacity, together 
with the necessary equipment for pouring the steel, and the 
accompanying buildings, total, according to Borchers, about 
$40,000 to $60,000. An installation with a 10- to 12^-ton 
furnace and a reserve furnace of the same size will cost about 
$60,000 to $80,000. 

The power consumption with the Girod furnace is about the 
same as that given for the Heroult furnace. What differences 
there may be due to a more or less favorable efficiency can be 
omitted when making arithmetical calculations, as the power 
consumption figures depend largely on the efficiency of the 
furnace, as well as on the charge and the final product. The 
composition of the final product produces much greater variations 
in the power consumption, than the differences in the efficiency. 
This, of course, does not hinder the furnace with the better 
efficiency to operate with less power and consequently with 
lower current costs, provided that an equal start is made with 
like raw materials, and like final products achieved. 

The electrode consumption with the Girod furnace may be 
taken to be the same as with the Heroult furnace, for there is 
no reason why the electrode consumption should be less with 
one furnace than with the other, when about the same electrode 


FIG. 650. 

FIG. 656. 


cross-sections are used in either case. Should there be given, 
nevertheless, larger or smaller values for the consumption 
figures, in one case or another, the larger wear can, in no case, 
be based on the principle of the furnace. Consequently if one 
furnace is to have any advantage over the other, it must depend 
on its more or less successfully constructed details. 

In order to give the reader an idea what these furnaces look 
like, Figs. 6$a and 6$b are shown, for these picture a 12^2- ton 
furnace as it is operated at Ugine, France. 

As the preceding chapter on the Heroult furnace was closed 
with Heroult's own opinion of the advantages of his furnace, so 
this chapter is closed with the deduction of Borchers, where 
he proves the superiority of the Girod furnace over the Heroult 
furnace. The quotation is taken from Stahl und Eisen, 
1909, page 1947, where Borchers says: "I strictly maintain 
that today there is no electric furnace for the refining of metal 
which excels the Girod furnace. I make special reference to the 
uniformity of the current distribution; the uniformity of the 
heat generation in the bath; the low voltage between poles, the 
consequent lesser insulation difficulties; followed by the con- 
sequent lesser danger to the operatives; on account of these 
circumstances, it excels in its simplicity of construction as a 
whole, and in its operation." 

It is well to compare this with the opinion of Heroult given 
on page 142. Lastly we may add that, if we consider only the 
evenness of the current distribution, and the heat generation as 
above mentioned, these alone should be enough to decide the 
question. That there is an advantage in the lower voltage goes 
without saying. To these we might add the .further advantages 
of the smaller current fluctuations, especially when melting down 
cold stock, while the opinion regarding the greater simplicity and 
the greater safety during the operation of one furnace over the 
other, may be left to the reader. Regarding the application of 
the Girod furnace, reference is had to the statistics in the 
closing chapter. Licenses for Girod furnaces may be had from 
the inventor, Paul Girod, Ugine, Savoy, France, or from his 
American representative, C. W. Leavitt, New York. 



IT was demonstrated in Chapter IV that an insulated wire 
of a coil carrying current generates lines of force, and that these 
lines or fields of force, are continually alternating, when alternat- 
ing current flows in the coil. These alternating lines of force 
constitute the well-known underlying principle for all induction 
phenomena. It is therefore evident that in an electrical con- 
ductor which lies in the field of another conductor, a current will 
be induced, which will be proportional to the number of lines of 
force cut in unit time. 

This fact immediately gives us the information, by the aid of 
which we are enabled to obtain any current strength by induction. 
We merely have to oversee that the conductor in which we desire 
to induce the current shall be cut with as many lines of force in 
unit time, as will give the wished-for current conditions. 

In order to achieve this we encounter these various possi- 
bilities : 

Imagine a certain number of lines of force, raised to twice 
their strength. Then we should find that a turn of wire, lying 
in this magnetic field, would have twice the electro-motive force 
generated in it as in a field of only the original strength. When 
the magnetic lines are doubled, then, the conductor is cut with 
twice the number of lines of force in the same time. 

The same effect is accomplished, however, when the field is 
kept at its original strength, if two turns are used instead of one, 
where they are both cut by the same number of lines of force. 
What we have then in this case is an increasing number of turns, 
and with it a raise in the voltage in the induced coil; because 
for the moment we may think of these two turns as being sepa- 
rated in such a way, so as to give us two separate turns, each 

1 60 


having the same voltage that one turn has now. Finally the 
potential in the induced circuit may be increased, by jaising the 
velocity of the current alternations, and this leads us to a change 
in the frequency. And as the induced voltage is proportional 
to the velocity of the alternating lines of force, it is evident that, 
a current of 50 cycles will give twice the induced voltage a 
current of 25 cycles will give, other things being equal. 

If we now combine the three methods into a formula, which 
influence the conditions in an induced circuit, we obtain 

e = CXvXsXN 

where e denotes the voltage 

v denotes the frequency 

s denotes the number of turns 

N denotes the number of lines of force 

and C is a constant. 

We have so far assumed that our lines of force, generated 
by the aid of a wire coil, sought their paths through the air. 
This arrangement is, however, very disadvantageous because 
the air is a very poor magnetic conductor (being only 1/180 as 
good as iron). The lines of force in this way seek the shortest 
path, resulting in the consequences (for instance, with a coil of 
a great number of turns) that only a part of the turns are cut 
by the total number of lines of force, whereas for the remaining 
turns only a part of the total lines are taken into consideration 
at all. In order to keep the lines of force from spreading, or 
straying, as it is called, we provide a good magnetic conductor 
for them, which forces them to take advantageous and prede- 
termined paths, due to the high magnetic conductivity, which 
in turn gives a good inductive action. These things give us 
the so-called transformer. 

Fig. 66 shows the principal arrangement of a transformer 
as it is commonly used, as well as for induction furnaces. In the 
figure, KI and K 2 denote the transformer cores, and J\ and J 2 
the yokes. The wire coils are wound on these cores. The coil 
receiving the current from an outside source is called the primary, 
and the coil delivering the useful current is called the secondary 


winding. Both coils are separated from each other by suitable 

If these yokes and cores were made from solid pieces of iron, 
then it would not be possible to avoid the considerate losses due 
to eddy currents, as set forth in Chapter IV. Therefore, in 
order to bring these losses down to the smallest percentage, the 
iron cores and yokes are built up of many thin sheets of iron of 

FIG. 66. 

FIG. 67. 

.3 to .5 mm. (.012 to .02 inches) thick. These sheets are insulated 
from each other by pasting sheets of paper on one side, about 
i/io as thick as the sheet iron, and the whole then held together 
by means of screws. Large core cross-sections are divided into 
separate divisions, which are kept apart by so-called ventilating 
ducts, by means of which the already low hysteresis and eddy 
current losses and their consequent heat generation are nullified. 
Fig. 67 shows one of these core cross-sections. 

If the primary coil of a transformer is energized with an 
alternating current, which must necessarily produce an induced 
current in the closed secondary circuit, then the iron core will be 
permeated with magnetic lines of force, which is common to both 
coils. As the primary and secondary coils, besides this, must 
have the same frequency, we obtain the equations for the volt- 
ages in both coils, as follows: 



from which it follows that: 

In this ratio we call the factor the ratio of transformation. 


The equation signifies that: 

The voltage is proportional to the number of turns. 

By applying a different number of turns in a transformer, 
we obtain a means whereby any existing voltage may be changed 
into any other voltage, and one thus suitable for the operation of 
electric induction furnaces. 

In this way transformers are nearly always used in alternating 
current installations. For this method makes it possible to 
transmit power over great distances at high voltages and at 
small currents, thus using only smaller and cheaper conductor 
cross-sections, from the central station to the point of power 
consumption. At that place then a transformer is erected, by 
the aid of which, the high primary voltage is changed to any 
desired secondary voltage, which may be most advantageous 
for the particular apparatus. 

We have already observed that transformers are used in this 
way for arc furnace installations. Alternating current provides 
such a convenient way of transforming energy in stationary 
transformers, and this together with its lack of chemical influence 
constitute the two factors responsible for the reason that all arc 
furnaces are operated with alternating current to-day. 

With the present technical perfection of the transformer 
this last may be regarded as a sort of interposed apparatus, 
which produces at a different voltage, almost the same amount 
of energy which it receives. That is to say, the losses in a trans- 
former are extraordinarily small. With transformers of more 
than 50 kw the losses are from 2 to a maximum of 3%. Even 
though the efficiencies of transformers for electric furnaces will 
fall slightly on account of the necessary overload capacity, yet 
we may consider, for the sake of simplicity, that the total primary 
power is given up in the secondary circuit. 


Then the primary power pi = e\ ii, 

and the secondary power p 2 = e 2 it, 

where pi = p 2 and consequently e\ i\ = e% iz- 

From this it follows that 

ii 2 Si , 

-r- = -- = that is: 

t 2 e\ Si 

The current is inversely proportional to the voltage and inversely 
proportional to the number of turns. 

The foregoing conclusion is of the greatest importance for it 
solves the building problems of induction furnaces. Induction 

furnaces in reality are 
nothing more nor less than 
properly designed special 
transformers. Hence every 
induction furnace has its 
iron core and yoke, to 
carry the lines of force, 
and a primary winding, 
wound over one part or 
another of the iron core. 
FIG. 68. FIG. 69. On the other hand, the 

secondary winding is com- 
posed either entirely or for the most part of the bath itself. 

This point of view enables us to group electric induction 
furnaces on the one hand into those furnaces where the 
secondary winding is composed entirely of the bath; and on the 
other hand into those where, besides the bath being the secondary 
winding, there is still another winding, made of copper - to aid 
the heating. We denote the former as simple induction furnaces 
and the latter as combination furnaces. 

If we take up the first group of simple induction furnaces, 
we see that the different methods of construction can be dis- 
tinguished merely by the way the primary coil is placed, relative 
to the bath. The Figs. 68 to 72 show a number of the most 
prevalent suggestions. In the figures the steel bath is 
denoted by the solid black, (the layer of slag is not shown,) 


the refractories by inclined hatching and the primary winding 
by cross hatching. Figs. 68 and 69 show the primary winding 
in the form of large radial disks, which are under or over the bath, 
or as Fig. 68 shows it to be both under and over the metal. On 
the other hand, Figs. 70 and 72 shows the primary winding in the 
form of a long cylinder, which is placed inside or outside of the 
ring-shaped hearth. With this arrangement we speak of-^ 
transformer with cylinder or tube winding and those of Figs. ( 
and 69 as having a disk winding. 

In all cases the principle of transforming the energy is the 
same, and in all cases we shall find the ring form hearth, in whose 
contents the heating currents are produced by means of induction, 
quite independent of the place in the magnetic circuit, occupied 
by the primary winding. It is evident that any of these winding 
schemes can be combined with every other method, and we may 
therefore state that there is no combination of windings and 
no placing of it at some part of the transformer, that has not 
already been patented as being particularly good. 

It has been shown that we are enabled to obtain any desired 
current strength in the secondary circuit, by properly winding 
the primary. The first one to recognize these conditions and 
use them in the design of an electric furnace was de Ferranti 1 , 

1 In this connection proper credit must also be given to Colby. Many 
years after the invention was made, the Franklin Institute investigated the 
early patent applications of both Ferranti and Colby and reported, in 1911, 
in part, in speaking of the patents, as follows: 

See British patent to Ferranti, No. 700, Dec. 16, 1887, filed January 15, 

U. S. Patent to Colby, No. 428,378, May 20, 1890, filed April 14, 1887. 

U. S. Patent to Colby, No. 428,379, May 20, 1890,- filed Sept. 19, 1887. . . . 

Between the years 1890 and 1900 no notable application of the process 
appears to have been made. . . . 

Colby's furnace is most broadly described in his U. S. Patent 428,379. 

It appears evident that the applicant was one of the first to devise the 
elemental features of the induction furnace. . . . 

It is generally conceded that the basic use of the transformer principle to 
electric furnaces was independently applied by both Ferranti and Colby, the 
dates of their patent applications being but a few months apart. The tubular 
water-cooled conductors, the means of supporting them and the connecting 
devices constitute essentially the features of novelty in the most recent patent 


who patented his apparatus, as shown schematically by Fig. 
68, in 1887. Even though his design was never put to practical 
use, we see how completely de Ferranti and Colby had at that 
time mastered the problem of heating by induction currents. 

If we use the furnace form as shown in Fig. 68, in order to 
obtain a clear view of induction heating, we observe that the 
middle core of the transformer carries the primary winding and 
that the furnace hearth is arranged concentric with this. There 

of Colby. . . . The forms of induction furnace depicted in the early 
Colby patents closely resemble those adopted in present-day apparatus and 
although but a joint pioneer in this field, his original designs are distinctive 
in anticipating the subsequent state of the art. 

In consideration of its originality and wide and successful commercial use, 
the Institute recommends to the Philadelphia Board of City Trusts the award 
of the John Scott Legacy Premium and Medal * to Edward Allen Colby of 
Newark, N. J., for his Induction Electric Furnace. 

Adopted at the Stated Meeting of May 3, 1911. 

(Signed) WALTER CLARK, President, 
R. R. OWENS, Secretary. 

GEO. A. HOADLEY, Chairman of the Committee on Science and the Arts. 

* Medal shown herewith: 

FIG. 6qa. Facsimile of medal awarded to Colby for his induction 
furnace by the Franklin Institute. 


is no secondary winding of copper such as we usually find with 
ordinary transformers. Should the ring-shaped hearth be filled 
with molten iron, as shown in the figure, we may regard this ring 
of iron as the secondary winding, which is composed of only one 
single turn. Induced currents will, therefore, appear in this 
iron ring, the same as they would in every other electrical con- 
ductor which lies in an alternating current magnetic field. 

As the iron ring comprises in itself one short-circuited turn 
or a short-circuit consequently all of the energy of the secondary 
circuit is transformed into heat, as the secondary current has 
to overcome the resistance of the iron bath. The heat quantity 
generated is proportional to i 2 r, that is, it is proportional to the 
product of the square of the current and the resistance. As the 
resistance of the iron bath may be regarded as being practically 
constant for a given charge, it is evident that any desired tem- 
perature may be obtained 1 by raising the current and, of course, 
first of all, by a proper choice of the primary turns; for the 
secondary turns with these furnaces are always equal to unity. 

Suppose we had an induction furnace, possessing 100 turns 
in its primary winding, and at a definite voltage of, say 1000, it 
took 100 amperes, we would obtain a secondary current value of 


^2 = 21 X = loo. = 10000 amperes. 

On the other hand, if we had a furnace wound with 120 
primary turns, and taking the same 100 amperes as before, but 
at a correspondingly changed voltage, we would obtain a current 

Si 120 

'h = i\ X = loo. - - = 12000 amperes, in the bath. 

Sz I ' , 

These examples show how the number of turns influences 
the secondary current, and consequently the attainable tem- 
perature of the bath. It is, therefore, the part of the furnace 
designer to so choose his proportions, that he may in any case 
reach the desired temperatures, for his particular case. 

1 See Am. Electro-Chemical Society, Sept., 1912, paper by C. H. Vom Baur 
on "Electric Induction Furnaces for Steel," giving an instance where the 
temperature of steel in an induction furnace reached 2550 to 2600 C. 


During the operation it is, of course, precluded that any 
primary turns of the furnace transformer be changed. Still, 
during the time of operation, temperature changes are desired, 
which in turn calls forth changes in the energy absorption of 
the furnace. But even these changes are easily made. We have 
only to realize that the load on the furnace transformer is brought 
about solely by the particular resistance of the iron bath, which 
we may consider as a constant factor for a definite charge. It 


is now apparent that we have in Ohm's Law i = , a simple 

remedy for changing the energy, and thereby the current, by 
simply altering the voltage for the primary winding. 

Necessarily the secondary voltage and its current are instantly 

changed as e 2 = e\ . 


If we now review the above, regarding induction furnaces, we 

1. The charge in induction furnaces is heated, solely and 
.alone by reason of the current overcoming the opposed resistance, 
and to any practically desired temperature. The induction 
furnace is therefore only a particularly favorable type of resis- 
tance furnace, which allows a complete and even heating of the 
metal, without producing any overheating at any point. 

2. By changing the primary voltage at any time during the 
operation of the furnace, the temperature of the charge may 
be raised or lowered at will, either quickly or slowly. At the 
same time the heat in the entire furnace contents is altogether 
uniformly raised or lowered. 

If all induction furnaces possess these qualities, what differ- 
ences are there then between the different arrangement of the 
windings as far as the molten metal is concerned? (See Figs. 68 
to 72.) 

It is well to state that the differences between Figs. 71 and 
72 are purely constructive, as the double magnetic path halves 
the cross-section of Fig. 72, opposite the simple path with the 
whole cross-section of Fig. 71, yet does not in any way produce 
any new electrical effects. Therefore, these two types of Kjellin 



furnaces only differ in their outward appearance, without either 
one or the other of the iron cores giving any advantages worth 

Substantial differences may, therefore, only be found in the 
arrangement of the coils, and these follow different directions. 
It is evident that the suggestion of Colby, made in 1887, (the 

FIG. 70. 

FIG. 71. 

FIG. 72. 

first to surround the hearth with the winding,) necessitates more 
copper conductors than the second suggestion of Kjellin, of 1900, 
where the primary winding is inside of the ring-shaped hearth. 
The whole arrangement of the Kjellin furnace, by reversing this 
idea, is simpler than the Colby furnace, not only on paper, but 
also in reality. The Colby furnace, as well as the de Ferranti 
furnace, are today only of historical importance, except for their 
later existing patents. This leaves only the accomplishments 
of Kjellin and Frick for discussion. 

If we put aside for the moment the fact that the Frick furnace 
does not permit such a general view of the hearth, or allow the 
accessibility thereto, on account of the overhanging disk winding, 
as we have with the Kjellin furnace with its coil removed from 
the operating conditions, we find that the chief distinction be- 
tween these furnaces lies in the different circulation of the bath, 
caused by the changed position of the coil. We saw in Chapter 
III that the motor effect of an electric current appears, when two 
conductors with their magnetic fields mutually affect each other. 
The different position of the winding cannot, therefore, be with- 
out its influence on the inclination of the bath surface. As will 


be shown in the following chapter, the Kjellin furnace produces 
the effect of pressing the molten metal toward the outside, so 
that it stands higher on the outside wall than on the inside. In 
the Frick furnace, for the same reason, we find a stronger mag- 
netic pressure on the current carrying molten metal, where the 
bath and the coil are nearest to each other, and this causes the 
metal surface to be more depressed at this point than the re- 
maining part of the hearth. The Frick furnace, therefore, also 
has an inclination to its bath surface, so that this stands higher 
at the outside than at the inside. While this slope in the bath 
is only 4 34' to 5 5' with an 8- ton Frick furnace, according to 
the published report of von Neumann of the firm of Freidrich 
Krupp (see Stahl und Risen, 1910, p. 1071), we see that 
with a Kjellin furnace of the same size, that it is 34. These 
differences naturally cause considerable deviation in the circula- 
tion phenomenon of the bath, so that these are greater in the 
Kjellin furnace and to its detriment, than they are in the Frick 

Even though there are certain differences between the Frick 
and the Kjellin furnaces, owing to the different position of the 
windings, still in the essentials of their operation they are en- 
tirely alike. As the Kjellin furnace opposite the Frick furnace 
has found a much more extended use, it will suffice if we describe 
the Kjellin furnace in the next chapter as a representative one. 
In this the secondary coil is composed solely and alone by the 
hearth metal itself. The honor is due Kjellin for producing the 
first practically useful induction furnace. 

In addition to the group of induction furnaces just men- 
tioned, in which the secondary coil of the furnace transformer 
is composed entirely by the bath, there is yet a second group 
of induction furnaces, which has another common copper 
winding, besides the short circuited secondary turn which is the 

This second group of induction furnaces owes its existence 
primarily to the fact that the furnaces of the first group have a 
comparatively poor power factor. The cause of this being that 
the distance between the primary and secondary windings is so 


great, that a large number of the lines of force take their path 
through the air, without being able to affect the secondary volt- 
age. We designate these lines of force as, stray lines or leakage 
lines, and the phenomena itself is called magnetic leakage, and 
it is this which operates heavily against the power factor. 

The greater the distance between the primary and secondary 
winding, the larger the magnetic leakage will be, and the lower 
the power factor. The leakage may be lessened by placing 
conductors in the path of these leakage lines, in which secondary 
currents are generated by induction. As these currents, which 
are generated by induced currents, always have the opposite 
direction of the primary or incoming current, (as was shown in 
discussing the self-induction phenomenon on page 44,) they 
will in turn send out stray lines of force in the opposite direction 
into the original stray field, and in this stray field the conductors 
lie. We may look upon this effect as one where the stray lines 
are pushed back, and in this way the power factor is raised by 
the coils, which lie in the space between the primary winding 
and the bath. 

Patents show a 'large number of suggestions, in which second- 
ary copper windings are to be employed, in order to gain the 
above result. But the fact must not be overlooked that on 
the one hand a poor power factor increases the initial cost but 
does not increase the energy losses, and on the other hand the 
current generation in the secondary winding (to decrease the 
magnetic leakage) can only be accomplished with energy so it 
is immediately evident that danger lurks nigh, in cur'ng a small 
evil with a larger one. This error is shown by all the designs, 
whose sole object it is to lessen the stray fields, by means of the 
secondary copper winding, in which the heat which is generated 
in these coils is not put to any use; on the contrary this results 
in only enlarging the cooling appliances for the windings, in 
order to protect them from too high a temperature. 

The idea of applying the above-mentioned stray field reducing 
arrangements to induction furnaces, can hardly be looked upon 
as induction furnace improvements, (as we have learned to know 
the furnace in the first group,) as long as no provision is made 


to profitably use the currents generated in the secondary copper 

This last requirement is fulfilled by the Roechling-Roden- 
hauser furnace, and, as a result, these furnaces have already come 
into quite extensive use, whereas all other suggestions to improve 
the induction furnace power factors from within the confines of 
the furnace proper are today only on paper. 

It therefore seems sufficient within the limits of this book, 
besides describing the Kjellin furnace, to merely narrate the 
details of the Roechling-Rodenhauser furnace, not only because 
these two are the only induction furnaces having found ex- 
tensive use, but also because a discussion of these furnaces will 
be sufficient to give the reader a clear idea of the workings of 
induction furnaces. At the same time it will also enable one to 
adequately judge the value of any other constructional features. 

If in closing we again mention the essential thing about an 
induction furnace, we find that the characteristic mark of them 
all is the common transformer. 

In the induction furnace we find the application of the 
built-in transformer to be of the greatest importance to the 
heating method. For in this way only is it possible to generate 
the strongest currents directly in the iron bath without con- 
duct or losses, so that the molten metal itself may be regarded 
as the source of heat. 

In his addresses before the "Verein deutscher Eisenhutten- 
leute," Borchers says: 

"Here in the induction furnace we should truly possess the 
most perfect of electrical heating. Here the generation of heat 
goes on solely and alone in the metal to be melted, and in the 
molten bath: the heat transference from other heat sources to 
the metal is not first required." 

Again when comparing resistance furnaces, and the induc- 
tion furnace may be regarded as a resistance furnace, with arc 
furnaces, Borchers says: "With both furnaces it is possible to 
reach a temperature of 3500 C. (6332 F.). There will always 
be 3500 C. at the arc of an arc furnace, while resistance heating 
enables any temperature up to this point to be reached." 



THE first induction furnace which made a name for itself 
as a result of its achievements was the Kjellin furnace. It 
was conceived in 1899; thus the first trial furnace was placed 
in operation on March 18, 1900. The furnace was only intended 
for a capacity of 80 kg. (176 Ibs.), with an energy consumption 
of 78 kw. Steel castings could be made with this furnace, 
only with the extraordinarily high power consumption of over 
7000 kw hours per ton of steel. With the second furnace of 
1 80 kg. (about 400 Ibs.), which was ready for operation in 
November, 1900, this amount was reduced to one- third of the 
original figure. A third furnace followed having a capacity of 
1350 (about 3000 Ibs.) to 1800 kg. (about 4000 Ibs.), which was 
installed in Gysinge on the Dalelf in Sweden. With this furnace 
they succeeded in bringing down the power consumption to 
about 800 kw.-hrs. per ton when making steel from cold scrap, 
and thus, the Kjellin furnace proved its practical and economical 

On account of the successful operation of these furnaces, the 
Kjellin patents were acquired by Siemens & Halske A.-G.- for 
the principal countries of Europe, and under their guidance 
these furnaces were soon used to a considerable extent. 

In the construction of the Kjellin furnace, the part giving it 
its characteristic appearance is the transformer, which comes 
up through the centre of the ring-formed hearth. The first 
successfully useful Kjellin furnace was the one having a capacity 
of 1350 to 1800 kg. (about 3/4000 lb.). This furnace is shown 
in its later design in Figs. 73 and 74. The original is of the 
stationary type. The transformer consists of two vertical cores 
and two horizontal yokes. These are composed of thin iron 



sheets, paper insulated, of the usual transformer construction, 
so that the magnetic losses are as low as possible. Whereas the 
yokes and the unwound core of the transformer have a rec- 


FIG. 73. 

tangular cross-section, the core carrying the primary winding is 
made in the form of a cross (see Figs. 73 and 84). This arrange- 
ment permits, on the one hand, a saving in the copper winding, 

FIG. 74. 


as the core is of circular form and easily wound, on the other, it 
provides for successful cooling of the transformer iron on account 
of its larger surface and thus favorable cooling conditions are 
provided. For this Kjellin with his first i^-ton furnace, used 
four one-inch tubes which were placed, one each in the recesses 
made by the section of cross form. These tubes carried an air 
circulation of 40 mm. (1.6 inches) water gauge pressure in this 
winding space, which was thus kept at permissible temperature. 
This air cooling was also taken from the normal transformer 
design and utilized in this special construction of furnace trans- 
former. Besides this, in order to shield the transformer, and 
especially the coil from the radiated heat, (from the furnace 
refractories,) the latter is surrounded with a double walled 
cylinder of brass of i> mm. (.06 inches) thickness. Either 
cooling water or air is passed through this protective cooling, in 
order to keep the heat from the winding and the transformer. 
The temperature of the cooling water coming from the pro- 
tective cylinder was measured during operation and showed 40 
to 50 C. Naturally this protective cylinder could not be a closed 
circuit, or if so, it would form a short circuited turn, which would 
become heated or even melted under the influence of the currents 
which would be induced in it. In order to avoid this the protect- 
ive shield is built as an open double walled ring, while in the 
Kjellin furnace it is bridged over with wood insulation. On 
the outside of these cylinders we find the furnace refractories 
or the brick work, in which there is a ring-shaped space concentric 
with the winding, which comprises the furnace hearth. The 
furnace shell is of sheet iron and encloses both cores of the trans- 

After the protective brick work has been placed in tne 
furnace, the bottom is rammed in. Then a templet having 
the shape of the hearth, is lowered into the furnace, so that the 
hearth walls of suitable material may be tamped in. When 
this work is finished the templet is raised and the hearth is 
practically ready. The hearth roof consists of special bricks, 
or of small refractory arches held in iron frames, so that they 
may be easily removed. This is necessary as the furnace has 


no doors, and the hearth and the progress of the charge can 
therefore only be watched by lifting off one or more covers. 
When the furnace is made of the stationary type, it must neces- 
sarily have the stationary type spout. 

Subsequent to the design of the first Kjellin furnace as just 
described, the following constructive changes were made: 

In order to allow of a thorough cooling of the transformer, it 
was divided into a number of smaller divisions, which were 
separated by means of suitable air spaces. This was only 
following good transformer practise, and the separate sheets 
were, of course, paper insulated as usual. The air cooling was 
changed so that there was a more uniform cooling, not only of 
the transformer iron, but also of the coil. 

The water cooling of the protective cylinder was avoided 
and air cooling substituted, this coming from the same ventilating 
fan feeding the coils. This simplified the furnace construction 
considerably, and gave equally safe operating conditions. 

The furnace was made of the tilting variety which materially 
bettered the conditions for teeming. It may be of interest to 
mention that Kjellin furnaces have been constructed as though 
they were self-heating pouring ladles, with which, for instance, 
the metal could be taken from the open hearth furnaces, then 
refined and finally poured from the furnace directly into the 
ingot moulds. 

The operation of the furnaces is primarily influenced by the 
fact that the molten metal serves as the secondary winding. 
Therefore as long as the metal does not possess a conductivity 
giving an operating voltage having a sufficient heating current, 
the heating of the furnace by electrical means is impossible. 
These conditions are the determining factors for the heating 
of the furnace. As the hearth is of the ring form, it is not 
feasible to heat the furnace with coke. Care is therefore taken 
with Kjellin and all other induction furnaces to heat them up 
with rings of material later to be melted. For making steel, 
these rings may be cast, welded, or even screwed together, and 
laid in the furnace. As soon as the current is turned on, induced 
currents arise in the iron rings, as they become short-circuited 


secondary windings. The iron is soon brought to redness, so 
that the heat thus produced can be used to warm up the furnace. 
As soon as the furnace walls are red hot, the furnace is charged 
with fluid metal, and the heating rings are subsequently melted. 
When this is accomplished, the furnace soon reaches the proper 
temperature so that the normal furnace operation may begin. 
If the furnace is operated with hot charges, as is often the case 
with Kjellin furnaces operating in conjunction with open hearth 
furnaces, the furnace is fully emptied after each charge and 
then charged again with open hearth metal. It is evident that 
it is possible to fully empty the furnace after each charge. Then 
when the fluid metal of the new charge, immediately makes a 
closed ring again, the heating begins simultaneously, provided 
the primary circuit is closed. 

The conditions are different when the furnace is charged 
with cold material. If, under these conditions, the furnace was 
completely emptied, and a ring made of a large number of 
pieces of cold scrap, its resistance would be so great, that the 
proper heating currents could not exist. In this case, it would 
be found useless to try to obtain a melt. It might, however, 
be possible to raise the secondary or bath voltage sufficiently, 
so that arcs would appear between the many small pieces of 
scrap. In such an event we would obtain heating methods 
similar to those employed when melting down cold scrap in the 
Girod furnace. The raising of the voltage necessary to do this, 
however, leads to difficulties in the transformer design. For 
this reason, therefore, when working with cold stock, a sufficient 
portion of the previous charge is left in the furnace to form a 
closed circuit. If the furnace is now further charged with scrap, 
it will be melted down by the heat generated in the metal from 
the previous charge. In this case the cross-section of the bath 
grows, and a greater absorption of energy takes place, thus 
hastening the melting. A very quiet melting together of the 
charge occurs in this way, without any sudden power fluctua- 
tions. As there is always a molten remainder in the furnace 
when using the method of cold charging, it is of advantage to 
keep this remainder as small as possible; still it must be large 


enough to render certain the closing of the molten secondary 
circuit. The smaller the section of the lower part of the trough, 
the easier it is to accomplish this. For this reason it is well 
to make the channel V-shaped. 

It was previously mentioned that cold charges are melted 
down without any current fluctuations taking place. 

We now come to the electrical conditions existing in the in- 
duction furnace. If we look a little closer at the current con- 
ditions of the Gysinge furnace, we find that for its operation 
there is provided a 300 HP water-wheel driving a direct connected 
165 to 170 kw, 15 cycle, single-phase generator of 3000 volts. 
When the furnace content is 1350 kg. (about 3000 Ib.) the 
power factor is 80%, and with a content of 1800 kg. (about 
4000 Ib.) it is 68%. 

Even these figures show the dependence of the power factor 
on the size of the charge with Kjellin furnaces. This is also 
substantiated by the curve in Fig. 77, which was made from 
results taken from a Kjellin furnace having a maximum capacity 
of S}4 tons. This shows, too, how (with other electrical con- 
ditions remaining the same), the power factor becomes lower 


_; ..0.55 

01234 56789 / 

FIG. 77. 

with an increased charge. We can, therefore, establish the 
fact that: "With the same frequency the power factor falls with 
an increased charge." 

In searching for the cause of this, we must go back to the 
causes affecting the power factor. For this purpose we again 
reproduce the vector diagram originally shown as Fig. 30 in 



Chapter IV. We see that the size of the angle depends on the 
resistance of the bath r and again upon the factors m and L. In 
our examples, in both of which the periodicity remains the same, 
the factor m, depending upon the latter, also remains unchanged. 
Therefore, only r and L remain 
as means for reducing the power e 

It is evident from the dia- 
gram that when the resistance r 
of the bath is reduced the angle 
becomes larger and the power 
factor, or cos </>, consequently de- 
creases. If the length of channel 
remains the same, but the cross- 
section of the bath changes, the re- 
sistance will change, because r = 

p and as the example showed, 


FiG. 78. 

that raising the charge from 1350 kg. (about 3000 Ib.) to 1800 
kg. (about 4000 Ib.), that is about 33%; and as the cross- 
section of the bath increased in like ratio, it becomes ap- 
parent why it is that the power factor falls with an increasing 
metal charge in the bath. 

Beside the resistance of the bath, however, the coefficient 
of self-induction has a noteworthy influence on the size of the 
power factor. It was shown in Chapter IV that the coefficient 
of self-induction depends upon the form and arrangement of 
the conductors. In order to give the reader an idea of this 
influence, it may be said that for conductors of ring form having 
a circular cross-section, the following formula for the coefficient 
of self-induction holds good: 


~ -(4 

L - 

8 D 

log nat , 8) . io-9 . 

Here D denotes the diameter of the wire coil, and d the diameter 
of the wire itself. 

This formula, however, is only strictly correct provided the 
conductor is not in the vicinity of any good magnetic conductors. 


However, it follows that the coefficient of self-induction primarily 
depends on the surface surrounded by the ring formed conductor, 
and that the coefficient of self-induction increases, the larger the 
surface becomes. Besides this the cross-section of the conductor 
also influences this factor, and as the formula shows, the co- 
efficient of self-induction becomes a little better with increasing 
cross-sections of the conductor. This latter influence, however, 
is too small to nullify the lowering of the power factor, occasioned 
by the lowered bath resistance when the cross-section of the 
bath is increased. The proof of this is plainly seen by the 
examples given. 

From what has just been said relative to the power factor 
it is apparent what the active causes are, and why Kjellin and 
similar furnaces had to be built with ever decreasing periodicities, 
for increasing capacities. We have just seen, that the power 
factor decreases when the charge is increased, due to the lesser 
resistance to the bath. As we saw in Chapter IV, the lowering 
of the power factor necessitates a greater current flow than it 
would have at a higher power factor, in case the furnace is to 
receive the same power, at a lower power factor and at the same 
voltage. Heavier currents, however, demand an increase in the 
copper cross-section of the primary winding, which in turn in- 
creases the needed space for winding the coil. To this must be 
added that with an increased capacity the energy absorbed by 
the furnace is naturally greater, so that the processes to be 
followed may not be unnecessarily expensive. This, too, 
necessitates the use of a larger copper conductor, and consequent- 
ly further increases the winding space. With the same thickness 
of the furnace refractories, this can only take place, however, 
when the diameter of the ring shaped hearth is increased; and, 
as we saw before, this causes a larger coefficient of self-induction, 
and with it a further decrease in the power factor. With an 
increasing charge, therefore, the power factor would drop very 
fast, and this would very quickly lead to impossible operating 
conditions with the frequency remaining the same. Fortunately 
by lowering the frequency, we have a means for meeting the 
lowering power factor. In order to recognize this, if we refer 



to Fig. 78 again, we see that the power factor is determined by 
the equation 


COS (f) = JT- 


as r and L are given by the bath conditions, it is only possible to 
influence the power factor by altering m, and the power factor 
will, of course, be larger, and so much better, the smaller the 
quantity m. This quantity m is determined by the equation 
m = 2 TT v where v equals the number of cycles per second. By 

00.411.52 33.84 5 6 7 88.59 
Tons Capacity 

FIG. 79. 

lowering the periodicity, the quantity m is reduced, and hence 
the value of the power factor is kept within reasonable limits 
for any particular size of furnace. Kjellin also availed himself 
of this means, and the curve shown by Fig. 79, which appeared 
in the Elektrotecknische Zeitschrift, in 1907, in an article 
by Englehardt, shows under what conditions the lowering of 
the frequency is desirable, with Kjellin furnaces of increasing 


size actually built and operated. The conditions here described 
are also, of course, more or less applicable for every other induc- 
tion furnace having a channel hearth. 

It is well to mention that the lowering of the periodicity is 
not always feasible as the normal frequencies are 25, 50 or 60. 
It is not possible to change from one frequency to another by 
means of stationary transformers, in the manner employed for 
voltage transformations. If it is desired, therefore, to connect 
to an existing power station having a higher frequency than 
would be favorable for the furnace operation, it will be necessary 
to employ a rotary transformer. In addition a low power factor 
necessitates comparatively large iron cross-sections in the 
generator as well as in the transformer, and consequently greater 
copper lengths for the windings, making a more expensive in- 
stallation. In order to give an idea of this, we may say that a 
generator of 25 cycles only costs half as much as one of equal 
capacity of five cycles. 

As electrodes are avoided with Kjellin and other induction 
furnaces, the regulating apparatus for the carbon electrodes 
themselves is not needed, so that the furnace does not require 
any movable parts, except the covers. With the absence of 
electrodes there are consequently no electrode losses, which 
leads us to the efficiency of the Kjellin furnace. As the furnace 
is merely a specially designed transformer, the only losses 
occurring are the ones usually prevalent in ordinary transformers. 
These losses are due to the iron losses, which are caused by the 
continually changing direction of the magnetization, and the 
copper losses in the windings. As the secondary winding in 
Kjellin furnaces is solely composed of the metal in the hearth, 
all the losses which ordinarily appear here, are used instead to 
advantage, because all electrical losses manifest themselves 
as heat, and in this case the generation of heat is what is desired. 
Losses, therefore, can only occur in the iron core and in the 
primary coils. The purely electrical losses of the induction 
furnace transformer hardly exceed those of the ordinary trans- 
former. On the whole, the electrical efficiency of the induction 
furnace is the best obtainable in any electric furnace. As a 



proof of the highest efficiency of induction furnaces, it may be 
said that, the most frictionless transposition of electric energy 
imaginable into heat takes place here, as the current generated 
in the secondary circuit, i.e., the induced currents in the iron 
bath, are generated at their point of origin and directly changed 
into heat. 

Induction furnaces may, therefore, be built for any existing 
voltage, for to generate the particularly high current in the 
bath is only dependent on a suitable number of turns in the 
primary winding. It was pointed out, for example, that the 
furnace at Gysinge, having a capacity of about 1500 kg. (3300 

FIG. 80. 

lb.), is operated at 3000 volts, its primary coil is arranged with 
295 turns, so that we have a secondary voltage corresponding to 


or about 10 volts. As it is possible to use any existing high 

potential current on the primary side, it is necessary that this 
part be shielded against coming into any possible contact with 
other conducting material. This is accomplished by completely 
encasing the furnace transformer, i.e., the transformer is built 
with a protecting shell, so that contact with any dangerous parts 
during the operation of the furnace is practically impossible. In 
addition to this the protective covering is connected with a 
copper conductor to the earth, or grounded so that in case 
any high potential current should strike the protective shield, 
it would immediately become harmless and flow to earth. Hence 


all danger to those operating the furnace is eliminated, and the 
best proof of the absolutely safe operation of the furnace, is the 
fact that thus far no operatives of induction furnaces have been 

As the operation of the induction furnace is usually not 
easily understood for non-electricians, the schematic connections 
of a Kjellin Induction furnace installation are shown in Fig. 80. 
In this figure the left half shows the electric installation at the 
central station, and the other side the actual furnace installation. 
The heavy lines between the central station and the furnace 
indicate the main high potential conductors. This high tension 
current is measured with instruments, by interposing so-called 
potential and current transformers between them and the 
main conductors so that the instruments only carry low voltage 
currents. The thin full lines indicate the necessary wiring for 
this. If we neglect for the moment the dotted lines, we see that 
the full lines of the circuit in the central station as well as at the 
furnace show instruments at either place, consisting of an 
ammeter A, a wattmeter B, and a voltmeter C, which must be 
watched during the operation of the furnace. D indicates the 
current transformers for measuring current, and E the potential 
transformers for measuring the voltage. In order to protect 
the instruments, the fuses F are inserted, whereas G represents 
an automatic release, which cuts out the main current when it 
is overloaded and thereby protects the generator. The generator 
itself is shown by H. At the furnace we see M which designates 
the sectors on which copper brushes rub, (similarly to those 
usually used on a motor,) in order that the furnace may receive 
its current and still remain in its tilted position. From the 
contact rails, the current is carried to the primary winding N, 
in well insulated conductors, which generate the induced cur- 
rents in the channel O, whose contents simultaneously act 
as the secondary winding. If we also mention the switch 
Pj Sit the furnace, which permits the current to be inter- 
rupted there, we have referred to all the apparatus of the oper- 
ating circuit. 

Of great importance is the regulating apparatus of an electric 


furnace, which permits the furnace to receive much or little 
energy, and thereby enables the furnace to be operated practical- 
ly. We saw previously with arc furnaces, that besides this 
apparatus, automatic regulators were also necessary, in order 
to smooth out the current fluctuations occasioned by the action 
of electrodes, and to keep a predetermined and constant amount 
of energy at the furnace. These regulators are wholly absent 
with induction furnaces, as sudden power fluctuations with 
induction furnaces are absolutely precluded. We have, there- 
fore, only to confine ourselves to the apparatus which is necessary 
to regulate the incoming energy, and for this it is quite sufficient 
to alter the primary voltage of the furnace. 

In order to easily change the voltage during the operation 
at any time, a handwheel rheostat, or regulator is placed at the 
central station, as well as at the furnace, by the aid of which the 
magnetizing or exciting current is varied at the alternator. In 
Fig. 80, / represents the exciter generator, the heavier dotted 
lines indicate the main wiring of the excitation circuit, and the 
lighter dotted line denotes the shunt circuit of the exciter gen- 
erator or exciter. At the furnace is the small regulator L, by 
the aid of which the field of the exciter is regulated with a very 
light current, whereas the regulating resistance K enables the 
main current of the exciter to be changed. In this way it is 
possible to regulate the voltage at the central station as well as 
at the furnace, and in order to keep both regulating platforms 
in communication with each other, it is usual to have them 
connected by means of loud-speaking telephones. 

After this discussion of typical Kjellin furnace switching 
methods, which are applicable also to other induction furnaces 
having special generators, we may turn to the comparison of 
the Kjellin furnace with the ideal furnace. That the Kjellin 
furnace requires special generators more than any other furnace 
discussed in detail, was seen when discussing the influence of 
the power factor; this is the reason for the abnormal frequencies, 
which, up till now, have been necessary for all induction furnaces 
having only the ring-shaped hearth. The operation of a Kjellin 
furnace with other than its own special generator, is not 


practicable and this increases the installation cost, and affects 
the obvious advantages of the furnace. 

Among the special advantages of the Kjellin furnace, as 
the typical representative of the pure induction furnaces, we 
may count first of all the absolute avoidance of any sudden or 
undesirable power fluctuations, which must be classified as 
unavoidable with arc furnaces having vertical or inclined elec- 
trodes. That there are no reasons for these sudden power 
changes with Kjellin furnaces we see when we realize, that with 
this pure resistance heating, sudden power changes could only 

FIG. 81. 

occur, with sudden heavy cross-sectional changes of the bath. 
This condition, though, is positively eliminated because the 
cross-section can only vary when charging or when tapping the 
furnace, and as these operations are always the function of a 
greater or lesser amount of time, it is evident that cross-sectional 
changes can only appear gradually during this time, and likewise 
the resistance changes and changes the current strength. This 
is proved from the operation of all induction furnaces. There 
is, however, a decided advantage in avoiding any sudden power 
changes, for it is evident that the generator required for an 
induction furnace needs to be just large enough to carry the 
maximum load required over a period of time ; whereas a genera- 
tor for an arc furnace would have to be more liberally propor- 
tioned, considering the heavy load fluctuations. The curve 


of Fig. 57 shows to what degree these power fluctuations occur, 
and it is interesting to compare this with the one shown by 
Fig. 81, which latter shows a Kjellin furnace under various 
methods of operation. These conditions naturally tend to 
cheapen the construction of the generator for the induction 
furnace, so that the greater cost occasioned by the generator of 
abnormal periodicity is at least compensated for to a certain 

It was seen when discussing the switching mechanism, that 
the regulation of the incoming energy of a KjelKn furnace is 
accomplished in the simplest way imaginable. It may, therefore, 
be well to point out again that the regulation of the energy of 
an induction furnace is accomplished in the most ideal way. 
For with an electric furnace, the same temperature is found 
throughout the whole bath, so that any change of the incoming 
energy alters its temperature gradually without in any way 
causing any overheating at any one spot, which is always to be 
dreaded under the electrode in arc furnaces. 

It has also been mentioned that the induction furnace un- 
doubtedly has the best attainable electrical efficiency of any 
electric furnace, because all electrode losses are avoided, and 
hence only the transformer losses come into play with induction 
furnaces, except when a special generator is used, and then only 
the primary copper losses and iron losses appear in the trans- 
former parts built into the furnace. Transformer losses are, 
however, present with nearly every arc furnace, thus a transform- 
er is almost invariably erected as closely as possible to the furnace. 

As the Kjellin furnace today is always built of the tilting 
variety, it fulfills one more demand of the ideal electric steel 
furnace. On the other hand, the Kjellin furnace cannot fulfill the 
demand which provides for an easily surveyed and accessible 
hearth and herein lies its great weakness as compared with other 
furnaces; its use is therefore restricted to that comparatively 
small field, in which the requirement of an easily surveyed and 
accessible hearth is immaterial. Such occasions may, however, 
occur where a furnace is intended to be a substitute for the 
crucible furnace, in which a very pure material is mixed in the 


bath, or in case the charge from other furnaces is merely to stand 
in the electric. The Kjellin furnace or any other induction 
furnace with a ring-shaped hearth, is found preferable to be 
used in this way. The advantage it has over the crucible, is 
that much larger homogeneous quantities of a desired quality 
can be obtained, whereas crucibles always have a very limited 
capacity; it is, difficult therefore, to produce large amounts of a 
predetermined and regular composition. Furthermore, the 
induction only needs a very limited number of operating men. 

FIG. 82. 

Finally, a considerable amount of money can be saved in the 
crucibles (from $5 to $8 in regenerative furnaces, but as high as 
$20 a ton in non-regenerative "pan system " oil burning furnaces). 
As a substitute for the above, the Kjellin furnace seems admirably 
suited. For most other classes of work, however, the furnace is 
unsuitable; because it is practically impossible to thoroughly 
remove the slag from the ring-shaped channel hearth, and thus 
avoid affecting the purifying process for the succeeding slag. 
This fact has been proved in practical work by many and ex- 
tensive tests. 

One of the next requirements of an ideal furnace is the 
adequate circulation of the bath by the aid of which the furnace 
will produce a thoroughly regular material. On account of the 
magnetic conditions of the Kjellin furnace the circulation of the 
hearth metal is almost perfect. The proof of this was first 
published by Englehardt in The Electrotechnische Zeitschrift, 
in 1907, and is shown schematically in Fig. 82. Here 0/ denotes 


the lines of force generated by the primary coil, which take their 
path through the transformer iron, fa" denotes those which 
are generated by the secondary winding or the bath, and take a 
path through the transformer iron in the opposite direction, so 
that we have resultant lines of force, denoted by 0. On account 
of the large distance, however, between the primary and second- 
ary coils, there must also necessarily be a number of stray lines 
of force, which find their path through the air. As far as these 
are generated by the primary winding, they are designated by 
</>/, and when generated by the currents flowing in the iron bath, 
they are designated by <j> 8 ". These lines of force play a very 
important part, as the iron and the molten metal offer a much 
lesser resistance to the lines of force than the air does, so that 
it may be assumed that a part of the secondary lines of force 
find their way through the molten metal. Both lines of force, 
</>/ and < s ", therefore, essentially flow in opposite directions, as 
they are generated by currents having opposite directions. (We 
saw in Chapter IV that the induced current, i.e., the current gen- 
erated by induction, is always in the opposite direction to the pri- 
mary current, which is the case here.) Fig. 82, however, shows that 
the opposite direction of these two lines of force circuits, makes the 
direction of current flow the same between the primary winding 
and the bath. It is a fact that lines of force of the same direction 
repel each other, hence forces must appear between the primary 
winding and the bath, which endeavor to repel the molten metal 
from the primary coil toward the outside. In Fig. 83 this force 
is denoted by P s . Besides this the force of gravity also operates 
on the bath. Both forces work at right angles to each other, 
causing a resultant force. Accordingly the surface of the bath 
inclines at right angles to this resultant, as is shown in Fig. 83. 

As a matter of fact the inclination of the bath surface can be 
more or less plainly seen with all Kjellin furnaces. 1 The flow of 
the metal is from the outside upper edge towards the inner lower 
one, which in this way provides an intimate mixing of the metal 

1 The inclination of the bath of an 8-ton Kjellin furnace is about 24 (See 
Stahl und Eisen, 1910, p. 1071). 


in the bath. One explanation of this flowing or rolling of the 
bath may be made, if we assume that the parts lying on the outer 
upper edge are subject to a greater cooling than the inside lower 
lying metal, so that the higher lying, cooler and consequently 
specifically heavier metal portions will tend to sink to the bottom, 
whereas the hot portions will rise. 

The circulation described has the advantage of the most 
thorough mixing of the whole furnace contents without mechani- 
cal aid. With large Kjellin furnaces operating at low frequencies, 
however, it has frequently happened that the inner wall of the 

FIG. 83. 

refractories is quickly destroyed by the circulation directed 
against it. It was only after decided efforts on the part of the 
Poldihiltte at Kladno, Austria, in applying the refractories in 
a special way that they were able to withstand these attacks, so 
that the lining, even with an 8-ton Kjellin furnace, now lasts six 
weeks, (164 heats,) when melting cold stock, 1 and 494 heats 
with hot charges. 

In the discussion of the Kjellin furnace circulation, it must 
be stated that the pinch effect mentioned in Chapter III does 
not come into play as long as the furnace is used for melting iron 
because the bath cross-sections in relation to the current density 
are too large when this furnace is thus used. The pinch effect 
could only be found if the cross-section should be especially 
narrowed at particular places. 

If it has been shown that the Kjellin furnace is only a sub- 
stitute for the crucible, still it may be said concerning the sizes 
this furnace has attained, that at present the Kjellin furnace 

1 See also Iron Age, Nov. 30, 1911, when refining hot metal. 



has a capacity of 8 tons of steel. One of these furnaces is operat- 
ing successfully at the works of Friedr. Krupp, at Essen, Germany. 
Fig. 84 shows the transformer of one of these 8-ton Kjellin 
furnaces for only five cycles. It is hardly advisable to build 
Kjellin furnaces of a larger size than this, first because the fre- 

FIG. 84. 

quency would have to be reduced still further, and secondly,, 
because it is to be feared that there would be difficulties with 
the durability of the refractories. 

Regarding the total efficiency of the furnace, we append the 
following : 

Several reports have been made by Englehardt on the 
Kjellin furnaces. In Stahl und Eisen, 1905, page 205, where 
he speaks of melting a charge consisting of 1/3 pig iron and 2/3 
scrap, he figured with a theoretical power consumption of 489 
Kw. hours per ton. If we compare the results obtained with the 
i. 5-ton Kjellin furnace, where with a mixture as above it took 


966 Kw. hours to melt a ton of steel over a six-hour melting 
period, and 800 Kw. hours during a four-hour period, this 
gives a total efficiency of 50% for the six-hour melting time, 
whereas Kjellin himself mentions 47%, and an efficiency of 60% 
for the four-hour melt. It is interesting to observe how the 
shortened melting time raises the attainable efficiency of a 
furnace. The reason for this is that on one hand, the radiation 
losses are prolonged for six hours, whereas on the other they only 
occur for four hours. While melting, therefore, it is advisable 
to operate with as high an incoming energy as possible, and to 
hasten the work to the greatest extent. The 60% efficiency 
with a four-hour charge is to be considered as most favorable, 
considering the small size of the furnace (\% tons). The 
attainment of such efficiencies, with Kjellin furnaces having 
such an unfavorable hearth, as far as radiation losses are con- 
cerned, is only possible because the electrical losses are at a 
minimum. In spite, however, of the assumed greater radiation 
losses of the Kjellin furnace as compared with the arc furnace, 
we find that the induction furnace always has a higher total 
efficiency than the arc furnace. This becomes even more appar- 
ent with larger furnaces. In the same article as above, Engle- 
hardt gives an efficiency of an 8-ton furnace corresponding to a 
power consumption of 590 Kw. hrs., when melting cold stock. 
This gives a total efficiency of about 80%. That this figure is 
attainable as a matter of fact is best proven by the practical 
operation, where with this ring-shaped hearth a power consump- 
tion of only 580 Kw. hrs. per ton of steel was attained by the use 
of suitable heat insulating covers. 

Regarding the application of this furnace, we refer to the 
statistics in the closing chapter. 

The sale of and giving licenses for Kjellin furnaces is handled 
by the Gesellschaft fur Elektrostahlanlagen in Berlin; in Eng- 
land and her colonies, except Canada, by the Grondal-Kjellin 
Co., London, and in the United States and Canada formerly 
by the American Electric Furnace Co., New York, and at 
present by Siemens and Halske A G , New York. 



ALTHOUGH we saw in the previous chapter that the Kjellin 
furnace and the induction furnace having a ring-shaped hearth, 
are inapplicable for many uses, and hence at a great disadvantage 
with the arc furnace, still the induction furnace has important 
advantages which must not be overlooked, especially where this 
furnace in its original form finds its best field; viz.: in the re- 
placement of the crucible furnace. These advantages include 
the absence of electrodes, and consequent saving in operating 
costs and also the avoidance of the risk of accidental impurities 
from the electrodes contaminating the bath, which latter is 
especially feared when making tool steel. The electrical effi- 
ciency attainable is also much higher. The absolutely steady 
furnace operation is almost ideal, and this steadiness is equally 
excellent for the central station. Finally, we may regard the 
uniform heating effect throughout the entire bath of the in- 
duction furnace together with its strong circulation, an advan- 
tage over the arc furnace, even though the experience thus 
far gained concerning the influence of the high temperature of 
the arc on the quality of the steel is not yet extensive enough 
to form a conclusive opinion on this point. 

Realizing the good points of the induction furnace referred 
to above, it was not long before efforts were made to retain the 
advantages of induction heating. For the disadvantages of the 
single ring hearth were clearly recognized. Later on pains were 
taken to alter the hearth in such a way that it would meet the 
demands of the metallurgist, and to produce thereby an induction 
furnace which would be equal to any refining work. It was 
recognized that if at the same time the operating conditions 
could be bettered (these as we have seen with the Kjellin furnace 
13 193 


necessitate the use of machines having unusual periodicities), 
then the induction furnace would be able to enter into successful 
competition with the arc furnace in any field. 

It was these considerations which led the Rochling Eisen 
und Stahlwerke at Volklingen on the Saar, Germany, to consider 
the problem of re-designing the induction furnace. This de- 
cision was reached only after it was clearly recognized that the 
single-ring channel induction furnace was not practical for refin- 
ing work. 

The first German Rochling-Rodenhauser patent was applied 
for on May 6, 1906, and subsequently granted (No. 199354 *). 

FIG. 85. 

This patent covers an induction furnace as shown schematically 
in Fig. 85. It may be seen that both cores of the transformer 
are provided with coils, in contradistinction to the Kjellin con- 
struction. Both cores are surrounded by an induction channel, 
which are joined between the cores in the middle, forming a 
roomy working hearth. This middle section is heated by means 
of an auxiliary current supplied by a secondary winding wound 
next to the primary winding, which has the marked advantage 
of reducing the stray field, and hence improves the power factor. 

1 Corresponding to U. S. patent No. 877739 f J an - 28 > 1908. 



The furnace principal shown in the sketch is known as the Roch- 
ling-Rodenhauser furnace. This furnace was investigated for 
its usefulness at the Rochling Iron & Steel Works from July to 
September, 1906, by means of a small test furnace holding 60 
kg. (132 lb.), and operated from a 50 cycle circuit. Fig. 86 shows 
this furnace at one stage of the tests with suspended electrodes 
composed partly of conductors of the second class, being used 
as a mode of utilizing the auxiliary current, notwithstanding 

FIG. 86. 

that the patent specification mentions conductors of the second 
class for transferring the current, which is the method exclusively 
employed today. 

The tests with the small furnace were later continued with a 
somewhat larger furnace, holding about 300 kg. (660 lb.). 

In the course of the development a furnace of about 500 to 
750 kg. (noo to 1650 lb.), was ordered by and constructed for 
the Richer Hiittenverein, which company was desirous also of 
investigating this form of furnace, knowing of the tests carried 
out at Volklingen. Until this time, the small furnaces were 


provided with covers which had to be lifted when charging, but 
the Eicher Hiittenverein furnace was the first one to be supplied 
with doors, thus simplifying the furnace operation. This 
arrangement may, at the same time, serve to indicate how, with 
the progress of developments, the work of refining in the furnace 
was transferred more and more from the channels to the main 
hearth, where it is carried on today exclusively, the channels 
serving only as heat carriers, without in any way accomplishing 
any metallurgical work. 

One of the determining factors in the further development 
of the furnace was due to the erection in the spring of 1907, at 
Volklingen, of an 8- ton Kjellin furnace which operated at only 
five cycles. The exhaustive tests carried on with the aid of this 
furnace, furnished convincing proof that the ring-shaped hearth 
was unsuitable for extensive refining, which was the goal of the 
Rochling Iron & Steel Works. On the other hand, these small 
test furnaces, above mentioned, gave the most favorable results 
in the refining of steel. Because of this a Rochling-Rodenhauser 
furnace was built and designed for the electric plant which had 
been installed to operate the five-cycle Kjellin furnace. This 
first large furnace had a capacity of about 3 tons and was placed 
in operation on June 22, 1907. It soon demonstrated the ad- 
vantages of the new furnace principle for large units. 

In order, however, to render this furnace system adaptable 
to all conditions, there was still one further step to take, i.e., 
to derive means to operate the furnace with polyphase current. 
For as long as it was not possible to use polyphase current 
directly in the induction furnace, the advantage of the induction 
furnace in its being able to be operated with any voltage that is 
available, would be of minor importance. The reason for this 
being that as it is only possible to operate the furnace with single 
phase current, it follows that the installation of a rotary trans- 
former would be necessary when obtaining power from a three- 
phase circuit. 

As early as 1907, therefore, the constructive features of a 
polyphase furnace were considered, and in February, 1908, the 
first polyphase Rochling-Rodenhauser furnace was placed in 



FlG. 86a. Two Phase R.-R. In- 
duction Furnace. 

operation. This was designed for 50 cycles and connected to the 
3 phase electric plant of the Rochling Iron & Steel Works. The 
application of the furnace to polyphase current was patented in 
all industrial countries. 

These short remarks show the development of the Rochling- 
Rodenhauser furnace, which can be obtained to-day not only for 

single phase current, but also 
for two and three phase cur- 
rent, for any convenient 
voltage and for normal fre- 

In its present form the 
Rochling - Rodenhauser fur- 
nace consists of a casing of 
strong sheet iron, which is 
supported by means of a 
semi-circular saddle and rack 
on rollers, thus allowing the 
furnace to tilt. The tilting 

may be accomplished in any way desired, but is usually done 
by means of an electric motor and suitable gearing. 

The furnace transformer is built into the shell. The upper 
yoke of the transformer is arranged to be easily removable, while 
the lower yoke and the cores are securely fastened by bolts, to 
the furnace casing, so that the transformer may stay securely in 
position even though the furnace is tilted 45. If we now turn 
to the furnace in its single phase form as shown in Figs. 87 to 
89, which indicates a 5-ton furnace operating at 15 cycles, 5000 
volts, we find two cores of somewhat long-drawn-out rectangular 
form. The cores are composed of a number of sections, which in 
turn are built up of paper-covered sheet iron of .3 mm. (.012 inch) 
thickness, the sections being separated from each other by the 
ventilating ducts H. Each core carries a primary winding A y 
and a secondary winding B. The primary winding is connected 
directly to the incoming voltage intended for the furnace, in the 
foregoing case, 5000 volts. The current is led to the windings 
by means of the usual high tension underground cable and thence 


to brushes, by the aid of which it is carried to a copper slip ring. 
The current is then led directly to the winding, by stationary 
conductors carried on high tension insulators. These methods 
of leading the current to the windings have the advantage that 
the heating may be continued, for instance, when the slag is being 
rabbled off, i.e., the furnace continues to receive its heat when 
in the tilted position. The secondary winding lies next to and 

FIG. 87. 

separated from the primary winding by an air space, which is 
both an insulating protection and a cooling chamber. The 
secondary winding is composed of heavy copper strips and 
carries very heavy currents at very low voltages. From this 
secondary winding, copper connections lead upwards from which 
the current is led to the poleplates E. These connecting pieces 
are represented by lines in Fig. 87.- 

The whole winding arrangement is surrounded by two 
cylinders of copper, brass or monel metal, which are separated 
from each other by an air space. Similarly there is an air space 



between the secondary winding and the inner cylinder. The 
inner cylinder is closed at the top by means of dust catchers in 
such a way, however, that the cooling air from the furnace trans- 
former may escape at the upper end with the least resistance. 
The method of supplying the cooling air is shown by means of the 
central air duct in Fig. 88, and the air direction is shown by the 
arrows. As a matter of fact this represents the method of 

applying the air supply for Rochling-Rodenhauser furnaces 
today, for such a centrally located movable 'duct underneath the 
furnace is provided from which the air is led to both cores. The 
air is so divided that the greater part takes its path upward 
through the winding space, thereby cooling the coils and the 
transformer cores, whereas a smaller part passes through the 
space between the two protective cylinders M , in order to keep 
the heat radiating from the brickwork, away from the whole 
transformer construction. 

Air is alone used for cooling from a blower usually of very 


low pressure. At Volklingen for instance with an 8-ton single 
phase furnace the blower pressure only corresponds to 40 mm. 
(1.6 inches) water gauge pressure. These cooling arrangements 
have given the most complete safety to the furnace during the 
past 4 years' continuous operation. 

In order that the unavoidable cooling of the furnace trans- 
former may not cause too great heating losses, much considera- 

FIG. 89. 

tion was given to the best possible heat protection for the hearth. 
On that account the outer protective cylinders are surrounded 
with a layer of granular material, which acts as a heat protector. 
On the outside of this follows the real refractory mass of either 
dolomite and tar or magnesite and tar. In order to obtain a 
hearth of the desired shape as shown in the figure, a wooden or 
cast iron templet is lowered into the furnace after the bottom 
has been rammed in, in a similar manner as with the Kjellin 
furnace. On the side of this templet, the hearth walls are 
tamped in, which when the templet has been removed leaves 


the necessary space for the molten metal. The cross-section 
a b of Fig. 87 plainly shows that the hearth really has the shape 
of an 8. It may be seen that the transformer cores are sur- 
rounded on the outside by the narrow channels C, while between 
the cores lies the true hearth and working chamber. Working 
doors are provided at both ends of the hearth, which makes this 
easily accessible and hence greatly lessens the necessary attend- 
ance at the furnace, as the entire roof of the furnace covering 
both hearth and channels remains stationary throughout the 
whole working period. The channels themselves are not intended 
for the metallurgical process, but they are of course necessary 
to provide the induced heating currents for the hearth. 

Concerning the hearth refractories, it may be mentioned 
that following the dolomite and tar outer hearth walls, there 
is provided a layer of coarse-grained heat insulating material, 
and that finally between this and the furnace shell is placed a 
ring of heat protecting brickwork. All parts of the roof covering 
are easily removable, in order that they may be easily lifted off 
and quickly replaced, in case a new lining is to be rammed in. 

It w r as remarked before that the copper secondary winding 
B leads to the poleplates E. These plates are imbedded in the 
lining, as shown in Figs. 87 and 89. They are made of soft cast 
steel, and have the largest possible surface on the side toward 
the hearth. Between the poleplates and the bath is the hearth 
wall, which as we have seen consists of dolomite and tar, so that 
the poleplate is protected against direct contact with the molten 

Mention has been made in previous chapters that the re- 
fractories used are conductors of the second class. That is to 
say, these materials, which are non-conductors at low tempera- 
tures, lose their resistance more and more with increasing tem- 
peratures, until finally they become comparatively good con- 
ductors at the temperatures which are prevalent in electric 
furnaces. This property of conductors of the second class is 
utilized in Rochling-Rodenhauser furnaces to carry the current 
from the secondary winding by means of the poleplates to the 
molten bath itself. In this manner that portion of the lining 


FIGS. 90 and 91. 



over the poleplate may be designated as the mass which transfers 
or conducts the current. 

At the beginning of the furnace heating this mass will act 
as a large resistance in the secondary circuit. This holds true 
as long as the furnace is heated up with iron rings, exactly as 
is done with KjelKn furnaces, and the same conditions exist 
when the furnace is charged with its first hot metal, so that at 
first we only have simple induction heating. As the furnace is 
further heated, the temperature also rises in the conductor of 

FIG. 92. 

the second class in front of the poleplate, and the resistance 
consequently drops under correct conditions, the secondary 
winding soon carries a considerable portion of the total energy 
of the furnace to the bath. With the 8-ton furnace at Volklingen, 
this result usually takes place in twelve hours. 

For the normal operation of the furnace there is therefore 
a double heating; first, we have the single induction heating 
in which case the ring formed parts of the hearth are to be 
designated as secondary circuits, and secondly the heating from 


the current of the copper coil secondary winding, which is carried 
to the bath via the poleplates and conductive lining. The 
arrows in Fig. 87 show both current plates of the secondary side, 
which plainly show that the total current flows in the same 
direction through the hearth. 

The conditions are much the same in three phase furnaces. 
These furnaces receive the three conductors from the three phase 
circuit on three cores, which are very similar to the ones of the 
single phase furnace, in that each core carries a primary and a 
secondary winding. Here also we find the coils surrounded with 
an inner and outer protecting cylinder, through which the 
ventilating air flows. In building these furnaces, it was the 
endeavor, of course, to provide a single roomy hearth, which 
could be easily surveyed and be easily accessible. Following 
these maxims the designs shown by Figs. 90 to 92 were evolved. 
The central hearth is consequently surrounded on three sides by 
the cores. Toward the outside, (corresponding to the arrange- 
ment of the single phase furnace,) these are encompassed about 
by induction channels, which join together and form the central 
hearth A . The yoke is often bent around in the form of a horse- 
shoe, by the aid of which the cores are connected at top and 

In order to make the hearth easily accessible and visible a 
door is fitted between each two cores, of which the one opposite 
the central cores is supplied with the tapping spout. The furnace 
is therefore emptied in that direction. There is a poleplate 
having two arms near each door toward the bath, which is pro- 
tected by the conductive lining from the bath, as in the single 
phase furnace. The arms of the poleplates are connected with 
one pole of the secondary copper winding, whereas the free ends 
of the other pole are connected together to the neutral point A 7 ", 
by means of the copper bar connections there shown. 

In order that there shall be no misconception about the 
current connections of a , Rochling-Rodenhauser furnace Fig. 
93 is given which shows the schematic diagram of a single phase. 
Similarly Fig. 94 shows the schematic diagram of a three phase 



In Fig. 93 both primary coils of the single phase furnace 
are shown connected in parallel. This also applies to the 
secondary coils, which are connected in parallel by the poleplates, 
between which the current flows through the bath. Fig. 94 
shows the primary winding for a three phase furnace and their 
neutral point N i . The heavier drawn secondary winding of the 
three cores has one end of each coil connected to the neutral 
point N 2, while the free ends are also here connected to the pole- 

FIG. 93. 

plates, between which the current-carrying lining and the bath 
are connected as heat-resisting material.- Besides this both 
figures show the hearth and channel limits. The conduits of a 
channel and hearth form a short circuited secondary, in which 
the heating currents are directly induced. 

The operating method resembles that of the open- hearth 
furnace, as a roomy working hearth is provided, and hence the 
conditions are present for successful refining work. If the 
furnaces are to be heated up, and hot metal is obtainable, this 
heating is accomplished similarly to the method used with 


Kjellin furnaces, i.e., iron rings for starting are laid in the furnace 
channels before the roof is replaced, which rings serve to bring 
the lining to a red heat under the action of the induced currents. 
When this is accomplished, hot metal is taken from any con- 
venient melting furnace and charged in the electric furnace and 
with this charging the starting or heat rings are melted, which 

FIG. 94. 

permits a quicker heating up of the furnace, as the cross-section 
is larger and consequently the energy supplied is greater. This 
heating permits the furnace to be placed in operation in the 
shortest time. It has, however, the disadvantage that it neces- 
sitates the use of molten metal from some other melting furnace. 
Such apparatus is often available and the molten metal "may be 
obtained from converters, open-hearth furnaces, cupolas, crucible 
pots or even blast furnaces, so that the disadvantage is seldom 
felt. It is more difficult, when a source of molten metal is not 
available, to start the electric furnace, and a melting furnace 



would therefore have to be furnished just for this purpose. In 
order to avoid this disadvantage, trials were made at the works 
of the Rochling Iron & Steel Co., with the object of starting up 
induction furnaces without the use of molten metal. This test 
produced satisfactory results, and the method is patented. 

In accordance with this method, the starting rings are solidly 
packed with pieces of scrap, steel turnings, etc., until the heating 
channels and the hearth are completely filled. After the roof is 
replaced, the current is switched on, and the heating rings soon 
become red hot under the action of the current. This assumes 
that with the furnace voltage remaining the same, the absorption 
of energy will rise. With a 
2-ton furnace it is pos- 
sible, for instance, in twelve 
hours, to render the entire 
furnace contents fluid, and 
the normal operation may 
then start. When heating 
up with hot metal about 
eight hours would be neces- 
sary in order to proceed 
with the normal furnace 

A normal heat with a 
Rochling-Rodenhauser fur- 
nace is much the same as 

with an arc furnace. The dephosphorizing usually occurs 
first, after which the slag is completely rabbled off, so that 
no deleterious material remains to delay the formation of the 
new slag for desulphurization. The removal of the slag 
occurs by rabbling through the doors. Of course it is im- 
possible to remove the slag from the channels, where they sur- 
round the cores, as these channels are quite unsuitable for this 
work. On this account they are permanently closed in by the 
furnace roof, as is the hearth itself. As the slag cannot be 
removed from the channels some provision must be made so 
that the slag is prevented from entering the channels. This is 

FIG. 95. 


now accomplished by placing fire-resisting bricks of magnesite 
or dolomite across the commencement of the channels in such a 
way that the iron bath in the hearth is i or 2 cm. (3/8" to ^") 
higher than the lower edge of the channel bridging bricks. In 
order to avoid as much as possible the heat losses occasioned 
by the bridging channel bricks coming into direct contact with 
the metal bath, a further normal roof of ordinary fire-brick is 
placed over them, which helps to lessen the heat losses by en- 
training a stationary layer of air. Fig. 95 shows these channel bridg- 
ing bricks, together with the refractories surrounding the channel. 
Mention should be made of the behavior of the Rochling- 
Rodenhauser furnaces when melting down scrap. So far it has 
not been possible to avoid the necessity of having in the furnace 


Cha r? e 









WOO kg 







6 Houis 








Hot, i 






, KJ 



4 Hours 

FIG. 96. 

a portion of the charge, which, as we have seen with the Kjellin 
furnace also, is necessary when working up scrap, in order to 
provide the necessary circuit" for the induced current, so that 
the scrap charged in the furnace may be melted down under the 
influence of the electric heating currents induced in the remaining 
portion of the previous charge. When scrap is to be melted 
down, therefore, and no fluid charge is at hand, the disadvantage 
consists in not being able to pour the entire charge. A certain 
percentage, say a quarter or a third of the entire contents, must 
remain in the furnace. The conditions are of course different 
when operating partly with a fluid charge, as for instance from 



a converter or an open-hearth furnace and partly with scrap. 
Then the conditions are similar to working only with a hot charge, 
so that there is no reason for leaving any of the charge in the 
furnace. With a mixed charge, therefore, the metal is fully 
teemed after each heat, after which some fluid metal is taken 
from some other furnace and poured into the electric furnace, 
which permits of the flow of the induced current. Thereupon 
the metal to be melted is charged gradually or at once, to such a 
degree that the cold and hot furnace contents at times reaches 
the roof. After this no attention is necessary until the entire 
contents is melted down. This takes place without the slight- 
est current disturbance, while the current and kilowatt curves 
rise slowly, as shown by the curves in Figs. 96 and 97, which 

FIG. 97. 

were taken from an 8-ton single phase furnace, and from a 
i. 5-ton three phase furnace, respectively, both at Volklingen. 
In the latter curve it is to be noticed that one division denotes 
45 amperes, also 50 volts, also 30 kw., also .1 for the power 
factor, and ten minutes. The crosses on the bottom line denote 
that 100 kg. of scrap were charged. 

It is also of importance from the standpoint of the practical 
operation of a furnace system that it is convenient to shut down 
the furnace for a limited time, for instance over Sunday. During 
such stops, a Rochling-Rodenhauser furnace is partly charged 
or even filled to capacity. It is then sealed up, after which the 
current is switched off and the furnace requires no further atten- 
tion. When it is desired to start up again, the current is switched 
on for several hours, and the furnace is thus heated up anew. 


When starting up in this way with the furnace fully charged, the- 
8-ton furnace at Volklingen after a 20-hour shut-down is ready 
for normal operation in about six hours. 

The method of operation of the electric furnace is exactly 
the same as with the Kjellin furnace, as far as the induction 
heating is concerned, which is generated by means of the primary 
coil, in the ring-shaped portion of the iron bath. This is applied 
in Rochling-Rodenhauser furnaces, twice on single phase, twice on 
two phase, see Fig. ioia, and three times on three phase furnaces. 
This heating therefore does not require any further explanation. 
On the other hand it is different with the secondary circuit, 
which is composed of tjie copper winding wound directly over 
the primary coil. The current then flows through the pole- 
plates, the current-carrying lining, and the metallic bath. 

The object of the secondary circuit is to raise the power 
factor, and to aid the heating of the furnace contents. It was 
seen that the low power factor of the Kjellin furnace, especially 
the low power factor of the larger sizes, led to the use of machines 
having very low frequencies, which materially increased the cost 
of installation. The reason for this decreasing power factor is 
found in the low bath resistance, together with the high coefficient 
of self induction, which was caused by the great distance between 
the coil and the bath. It was therefore necessary to investigate 
these causes, if the above-mentioned lowering of the power factor 
was to be avoided. 

In order to increase the bath resistance the long rectangular 
form of core was chosen, in place of the more circular shape used 
with the Kjellin furnace. Furthermore, by placing the winding 
on two or three cores, it became possible to materially decrease 
the inner periphery of the induced part of the bath, as compared 
to that of the Kjellin furnace. This brought about substantial 
advantages, so that the power factor with Rochling-Rodenhauser 
furnaces stays much higher than with Kjellin furnaces having 
equal capacities and equal frequencies, even during the heating 
up period, i.e., at a time when the poleplate circuits cannot yet 
do much work, because the current- carrying lining has too high 
.a resistance. In order to further decrease the leakage as much 


as possible, because of the comparatively large distance between 
the primary coil and the bath still remaining, use was made of 
electric conductors placed in the path of the stray lines of force. 
This expedient was mentioned when discussing induction furnaces 
in general. 

The conductors of Rochling-Rodenhauser furnaces, which ar& 
placed in the path of the magnetic leakage lines, are used there- 
fore as secondary copper coils, so that the currents 'produced 
by lowering the leakage field are used at the same time to heat 
the metal bath. The influence of this secondary coil is most 
important, and can best be shown by the fact that with the i^- 
ton, three phase, 50 cycle furnace operating at Volklingen, the 
power factor rose from 0.5% at the start to .8% and above, as 
the work of the pole plates increased. The secondary winding 
meanwhile takes up from 20% to a maximum of 30% of the 
total work of the furnace. 

By using the above expedients, which consist of, ist-the 
bath resistance being increased within practicable possible limits, 
2nd the coils being wound on two or three cores, and 3rd the 
secondary copper coils used to reduce the leakage field, it is 
possible to build Rochling-Rodenhauser furnaces for standard 
frequencies, viz., 25 (50 in Europe) and, in the case of very 
large units, for 15 cycles without the power factor falling below 
values found elsewhere. Polyphase furnaces of 3^2 tons and 
50 cycles are quite practical, whereas large sizes up to a maximum 
of 15 tons would have to be operated with 25 cycles, with poly- 
phase current. 

With the given conditions in the secondary circuit of the 
Rochling-Rodenhauser furnace, the conductor of the second 
class, which is placed in front of the poleplates must be made to 
conduct as soon as possible. This is accomplished primarily by 
giving the conductor as large a cross-section as possible, and 
making the current path as short as possible, so that the operation 
proceeds only with very low current densities. In order, how- 
ever, to force a current passage, and thereby provide as quick a 
heating up as possible, with the comparatively high resistance 
of the conductor of the second class, a higher voltage is used 


in the secondary circuit during the heating up period, than later, 
during the regular operation. A convenient expedient for 
accomplishing this is to alter the number of turns of the secondary 
winding; this may be accomplished by the use of a single throw- 
over switch. In this way, for instance, the 8-ton furnace at 
Volklingen is operated with 20 volts in the secondary circuit, 
during the heating up period, with the primary voltage remaining 
the same, whereas, subsequently, i.e., during the normal operation 
only, 10 volts is used. With this voltage it is possible to conduct 
several thousand amperes through the current-carrying lining 
into the molten metal. This naturally brings with it an increased 
heating effect, even though the main heating is accomplished 
with the currents directly induced in the bath; also, during 
normal operation the current encounters a resistance in its path, 
in the current-carrying mass leading to the bath, similar to the 
resistance mentioned with carbon electrodes in arc furnaces. 
In accordance with this, considerable portion of the energy 
generated in the secondary coil must necessarily be converted 
into heat in the current-carrying lining. This would, of course, 
mean substantial losses, provided this heat could not be utilized 
in the bath. This heat is however utilized as the metal bath is 
in contact with the current-carrying lining, which may be re- 
garded merely as a heat resistor, the object being of transferring 
the heat generated in it to the bath. Slight radiation losses only 
occur as a small percentage in the arms of the poleplates, 
(which however possess no cooling arrangements), but there 
are no appreciable losses otherwise, excepting those which for 
instance are occasioned by the radiation and heat conduction 
of the insulated heating channels, as already described. 

Considering the foregoing, we may regard the Rochling- 
Rodenhauser furnace as a combination of a pure induction 
furnace with a pure resistance furnace, so that the usual designa- 
tion of the furnace as a " combination " furnace is well founded. 

If the combination furnace be now compared with the ideal 
furnace, it may be said concerning the utilization of every available 
form of alternating current, that the furnaces fulfil this require- 
ment to a considerable extent, for they may be built for any 



prevailing voltage for either single or polyphase. A certain 
restriction, however, appears, in that the falling power factor 
with increasing size furnaces cannot be avoided, even though 
it is considerably less than with the Kjellin furnace. Single 
phase furnaces of 3- to 5-ton capacities are practical only for 25 
cycles and less; with greater capacities they can only be built 
for 15 cycles. With polyphase furnaces the drop is not so 
sensitive, so that here 3 -ton furnaces may be built for 50 cycles, 
and i5-ton furnaces may still be built for 25 cycles. 

From what has gone before, it is evident that sudden power 
fluctuations with Rochling-Rodenhauser furnaces are absolutely 
absent. Where value is placed on machinery having small 
repairs and long life, these 
furnaces accordingly mean an 
ideal load for the central 
power plant. This is also the 
case when it is necessary to 
change the energy supplied 
to the furnace, and thus raise 
the temperature to the degree 
necessary for favorable oper- 
ation of the metallurgical FIG. 98. 
process. If the furnace is to 

operate in conjunction with its own generator, it can best be 
regulated as shown by the wiring of Fig. 80 which is perfectly 
applicable to the single phase Rochling-Rodenhauser furnace. 
This method was originally applied at the Rochling works, to 
an 8-ton Kjellin furnace, and is used unchanged today for a 
furnace of the same size. In using this scheme it is assumed 
that the generator is to be used only for the furnace. As it has 
been shown that the combination furnaces have the advantage 
that they may be connected directly to existing polyphase cir- 
cuits of any voltage and frequency, even for furnaces of con- 
siderable size, this arrangement becomes particularly interesting. 
For instance the arrangement may be used at all works which do 
not desire to erect their own power plant, but wish to use current 
from a distant central station. It is, however, significant for 


works which desire to form a definite and practical opinion of 
the working operation of the Rochling-Rodenhauser furnaces 
to do so by means of a small trial installation. Such works 
would lay the greatest stress on the ability to utilize their existing 
power plant in order to reduce the initial cost of a trial installa- 

In such cases it is necessary to regulate the voltage at the 
furnace without appreciably disturbing the voltage of the 
central power plant. This is accomplished by the use of so- 
called regulating or auto- transformers. Fig. 98 shows the wiring 
scheme for one of these for three phase currents. If at the 
points, a 1; a. 2J and a 3 , for instance, a certain star connected 
potential of, say, 500 volts is connected, corresponding to a phase 
voltage of 289 volts, and if there are 289 turns between the 
neutral point and a lt a 2J and a 3 , then there are only 260 turns 
lying between the neutral point and the points &,, b 2 , and b 3 , 
there will be only a 260 phase voltage between these points and 
the neutral point, corresponding to a star connected voltage of 
260 X 1.73 = 450 volts. In case the primary coils of the 
furnace are connected with the points b lt b 2} and b 3 , we give 
them 450 volts in place of the 500 volts of the circuit. Any 
number of these taps may be brought out of the auto-transformer. 
For instance, the points c lt c 2 , and c 3 , could deliver 400 volts, 
provided it is assumed they correspond to about 230 turns so that 
a phase voltage of 230 would result. In the same way that a 
voltage decrease is attained a voltage increase may also be 
reached. In this way the points d lt d 2 , and d 3 would give 
550 volts, and the points e l} e 2y and e 3 , a potential of 600, 
if the number of turns per core were raised respectively to 318 
and 347. It may therefore be seen to be a matter of fact, that 
the voltages are proportional to the number of turns, and that 
only one continuous winding per core of the transformer can be 
used for voltage regulation. A so-called step switch, especially 
designed, is still necessary, in addition to this transformer, by 
the aid of which the winding may be switched from one point 
to another without interrupting the current. 

The electrical efficiency of a Rochling-Rodenhauser furnace 


may be regarded as extraordinarily favorable, for an electric 
furnace. Measurements taken on a 3^-ton single phase furnace 
in Volkingen, for instance, gave an efficiency of 96%, notwith- 
standing that this furnace was the first of the larger ones to be 
constructed, and could by no means be designated to be especially 
well dimensioned as the line losses are extremely low when 
using high potential directly, and as rotary transformer losses 
are usually not present, the total electrical efficiency of these 
furnaces will always be greater than 90%. 

It has already been mentioned that the furnaces are of the 
tilting variety. 

The requirements of an easily surveyed and accessible hearth 
may be regarded as being fulfilled, as the hearth is central and 
has two or three operating doors at the sides. There remain, of 
course, the heating channels at the sides which are not well 
esteemed by the metallurgist, although they are so arranged 
that slag cannot enter them, but it must be remembered that 
they result in a far-reaching circulation on account of electrical 
conditions, which assures a homogeneous composition of the 
molten metal both in the channels and in the hearth, so that the 
heating channels, as a matter of fact, exercise no deleterious 
influence on the operations. 

The circulation phenomena in Rochling-Rodenhauser furnaces 
result advantageously owing to electric and magnetic conditions. 
Referring to Fig. 99, which shows a hearth of a single phase or two 
phase furnace, the arrows show the direction of the circulation, 
which direction may easily be observed in actual practise by 
throwing some lime dust on the uncovered metal bath. We have 
also the circulation of the bath against the lining, between the 
bath and the coil. In addition to this it may be observed that 
the molten metal is somewhat elevated at the doors, which re- 
sults in a flow of the fluid mass toward the middle of the hearth 
on the one hand, and toward the channels on the other. Both 
manifestations may be defined as being the mildest forms of the 
pinch effect. This appears as shown in Chapter III, because 
the fluid conductor flows toward the point of higher temperature. 
The high current densities are to be found, first, in the middle oi 


the hearth, and second, in the heating channels, whereas the cur- 
rent densities are decidedly the lowest at the broad sides of the 
hearth near the doors. There is consequently a suction action, 
first on the part of the channels, and again on the part of the 
centre of the hearth. This circulation, based on the pinch effect, 
has the advantage that it works vertically against the inner lining, 
and therefore lessens this motion, so that throughout the whole 
furnace there can be observed only a slow flow, without being 
violent in any way. A part of the ascending motion of the 

FIG. 99. 

FIG. loo. 

fluid metal at the doors is to be accounted for by stronger 
heating of the bath, occasioned by the heat generated by 
the current through the current-carrying lining, which 
naturally results in a rise in temperature of the higher heated 

Exactly the same reasons cause the circulation phenomena in 
the three phase furnaces, so that there remains little to be said 
about it. Fig. 100 shows the circulation phenomena which may 
be observed in one of these furnaces. This is somewhat different 
from the single phase furnace, as there is an additional circular 
motion of the bath between the three cores. This rotary motion 
is the result of a rotating field, which arises between the three 
transformer cores and has a similar action to the connected 



stator of the polyphase motor, by means of which the armature 
is caused to revolve. This comparison must not lead one to the 
erroneous conclusion, that the bath rotates at the same speed 
as the rotor of a motor would under similar conditions. The 
motion is also very mild here, and can often only be observed by 

FIG. i oi. Transformer of a single phase furnace. 

throwing fine lime on the bare metal. A circulation, such as 
this, possesses distinct advantages for the metallurgical process. 
It causes new masses of metal to be brought into contact with 
the refining slag, also a thorough homogeneity of the contents 
of the furnace, and finally facilitates the separation of suspended 


particles of slag, without having any consequential disadvantages. 
If, now, the opinion is expressed that the lining does not stand 
up well under the action of the circulation, this may be accounted 
for by the fact that the durability of the lining of induction 
furnaces was short, as compared with that of arc furnaces; but 
the lining costs per ton of steel were not higher than those of 

arc furnaces. As a matter of 
fact the influence of the circu- 
lation of the molten metal on 
the lining is almost insignifi- 
cant, for the wear takes place 
only at the slag line and is 
therefore only to be account- 
ed for by the chemical action 
of the slag, which can be 
easily proved by the worn 
lining at the slag line. Fi- 
nally, it may be mentioned 
that the Rochling Iron & 

FIG. loia. Heavy lines denote the 
secondary circuit of a two phase 
R.-R. furnace with two separate 
pole plates at either end. 

Steel Works have been suc- 
cessful in constructing the 

refractory lining of Rochling-Rodenhauser furnaces to with- 
stand the action of the slag, to such an extent that the dur- 
ability of the hearth compares very favorably with that of 
the Girod or Stassano furnaces. 

Regarding the circulating phenomena, there is still to be 
mentioned that with very large furnaces, for instance, having 
comparatively great depths of bath, it may be advisable to 
obtain a stronger motion in the bath than is possible with the 
above-mentioned forces. A convenient means for doing this is 
to increase the pinch effect. In order that this may be accom- 
plished, all that is required is to raise the bottom of the hearth 
in the centre for a short length. This causes a contraction of 
the cross-section of the bath at this place, giving a higher density, 
and consequently a stronger suction action at the centre of the 
bath which can be increased until the bath becomes wavy, in 
case the raised portion is made high enough. By means of this 



arrangement, therefore, we have a convenient means of increasing 
the circulating motion to any desired degree. 

The Rochling-Rodenhauser furnace is far-reaching in its 
application. The furnaces are adapted to produce any quality 
of steel from any common raw material. The assurance for this 

FIG. 1 02. Transformer of a three phase furnace. 


is given by the similarity of the working hearth to that of the 
Siemens-Martin open-hearth furnace. It does not seem inap- 
propriate to dwell explicitly on this point at this place, as these 
furnaces, being a type of induction furnace, were credited with 
the same weaknesses that the early induction furnaces possessed. 
These disadvantages appeared with the Kjellin furnace (the 
first induction furnace which found its way into practical steel 
making) on account of its peculiarly shaped hearth. This 
prejudice against the furnaces is, however, entirely unjust. Re- 
garding this, reference is made to the second part of this book, 
where the best refining results attained with other electric 
furnaces, as well as with Rochling-Rodenhauser furnaces, are 
discussed in detail. Of course these furnaces reach a limit of 
their applicability, when cold stock is to be melted in the same 
furnace for high class steel alloys with quick changes following 
each other. In this case, when working up cold stock, the metal 
remaining in the hearth would interfere with the composition of 
the next charge. Therefore, if induction furnaces were to be 
used in this case, two furnaces would be necessary, one of which 
would be designated to melt the cold metal, and be operated to 
make a portion of each charge start the succeeding charge, 
while the second furnace, in which the refining and alloys would 
be made, could always be charged with hot metal from the first 
furnace, and would consequently be fully emptied after each 
charge. With this method of operation it is evident that the 
previous charge cannot in any way affect the quality of the 
succeeding one. It requires, however, a comparatively large 
initial capital, which would only be justified when it would be 
desired to make large quantities of electric steel. These con- 
ditions would make it difficult, if not impossible, for small steel 
plants to compete with the induction furnace in its present 
form, when using the above method. On the other hand, in 
very many other cases, the necessity of leaving some of the 
metal in the furnace can hardly be regarded as a detriment when 
working up scrap. This applies particularly to those making 
electric steel, in the manner it is made to-day, for instance, in 
large lots to take the place of Swedish iron. For, in this case, 


the metal remaining offers the advantage of allowing the melting 
operation to proceed by using a considerable proportion of the 
available electric energy left in the remaining metal, even while 
charging. This results in shortening the melting time, and 
produces a better efficiency and also a greater production. 

A further limitation of the use of electric furnaces may be 
ascertained by studying the limit of practicability of the furnaces 
according to their size. It may be mentioned here, that single 
phase and two phase furnaces are built for a minimum capacity 
of 300 kg. (660 lb.), and give practical and economically useful 
operating conditions. If the bath surface becomes too large 
in proportion to the capacity, then the thermal losses become of 
such an extent that an economical operation would no longer 
be possible. The useful limits of these furnaces for the iron 
industry lie therefore within the sizes mentioned above and 

The largest Rochling-Rodenhauser furnace unit so far built 
has a capacity of 8 to 10 tons. -This gives excellent operating 
results at the works of the Rochling Iron & Steel Works. This 
size furnace, though, in no way indicates the upper limit of 
its practicability or of the economical usefulness of the furnace. 
Complete constructional details have been worked out for 
furnaces of the 25-ton size, so that this size may be regarded as 
the upper limit, for the present, with which good operating results 
may be determined with certainty. 

The thermal efficiency of the furnace may best be judged by 
the total efficiency. That the efficiency of furnaces becomes 
better with increasing sizes, is true as it is with other furnaces 
previously discussed. We find that the smaller sizes, adapted 
to single phase, are considerably superior to the three phase, 
considering their total efficiency, whereas when the capacity 
reaches 3 tons the efficiencies are about equal, while for larger 
sizes than this the polyphase furnace is the better. The reason 
for this arises from the fact that single phase furnaces, even of 
the 3-ton size, must be operated from as low as 25-cycles, whereas 
three phase furnaces of this size may be operated to advantage 
with 50 cycles. Lowering the frequency necessitates enlarging 


the cross-section of the transformer, which means more space 
for the coils, and hence a larger periphery of the walls touching 
the bath surface, so that with a 3 -ton single phase furnace we 
have a larger bath surface, with a lesser bath depth, than with a 

FIG. 103. 

three phase furnace of the same size, which has a smaller bath 
surface with a greater depth of bath. 

Even though the single phase furnace up to the 3 -ton size is 
preferable to the three phase furnace on account of its efficiency, 
these conditions, however, become a deciding factor only when a 
new generator is to be furnished for the furnace in either case. If, 
on the other hand, an extensive power plant producing a certain 
type of current already exists, then the choice of furnaces 
would in most cases be decided by the actual current available, 
in case this could be used directly in a single, two, or three 
phase furnace. In these cases, when using the existing current 
directly in the furnace, the deciding factor would be the avoidance 
of the rotary transformer losses, which are always about 15 to 
20%, and therefore so large that, as fai as the total efficiency of 


a furnace installation is concerned, they would be bound to be 
the deciding factor. 

In discussing the question of efficiency, the following will be 
of interest: The total (net) efficiency of a ij^-ton three phase 
furnace at the Rochling Iron & Steel Works was 60%, when 
comparing the theoretical figures and the actual amount of 
energy used in melting up cold scrap. The total efficiency of 
the 8-ton furnace operating at Volklingen was determined by the 
fact that it took 580 kw. hrs. to melt one ton of common scrap. 
If we compare this with the required theoretical energy, which 
was placed at 489 kw. hrs. in the previous chapter, we find that 
the 8-ton Rochling-Rodenhauser single phase furnace has an 

efficiency of -~- = 85%. 

Even though these figures may not be called absolutely 
correct, on account of unavoidable irregularities or uncertainties 
creeping into the theoretical computations of the energy required, 
still the fact remains that the required power of 580 kw. hrs. was 
all that was needed to melt one ton of common commercial steel 
scrap, so that the efficiency figures retain their full accuracy and 
significance as relative figures of comparison. Considering the 
heat losses, these results show that, in spite of the really unfavor- 
able arrangement of the hearth with the side connecting channels, 
efficiencies are still attained, which are fully equal to those 
of the Kjellin furnace, with its ring formed hearth, and they may 
also be considered as comparing most favorably with the 
efficiency of any arc furnace. 

In adding a few words here on the installation cost, reference 
is made to a 5-ton polyphase furnace which is to be connected to 
an existing power plant. This would operate in conjunction 
with a separate transformer and a multi-point switch and would 
cost about $18,000. This price includes the furnace, the furnace 
transformer, the switchboard, the electrical tilting mechanism, 
etc. It, however, does not include the generator installation, 
which was assumed to be already in existence. 

The following references are to the figures which augment 
the text. Fig. 101 shows the transformer of a single phase 


Rochling-Rodenhauser furnace, while Fig. 102 shows the trans- 
former for a three phase furnace. From the figures one can 
plainly perceive the arrangement of the cores and yokes. The 
former are covered by the protecting cylinders, as these figures 
show the manner in which the ventilating air is conveyed by 
means of a central air duct, as shown in Fig. 88. The bifurcated 
air-supply duct which lies between this central duct and the cores 

FIG. 104. 

is also plainly distinguishable in Fig. 101. Such furnace trans- 
formers are then built directly into the furnace brickwork or into 
the furnace refractories, which thus decide the appearance of 
the furnace. Fig. 103 shows an 8-ton single phase furnace in 
its tilted position, and Fig. 104 a three phase furnace of i^ tons 

The sale of these furnaces and the giving of licenses are con- 
ducted in Continental Europe by the Gesellschaft fur Elektro- 
stahlanlagen, Berlin, Nonnendamm; for England and her Colonies 
except Canada by the Grondal Kjellin Co., London. In the 
United States and Canada formerly by the American Electric 
Furnace Co., New York, at present by Siemens and Halske, A 
G , New York. 


IN the consideration of electric furnaces that one must not 
be overlooked which may be briefly called the Electric Shaft 
Furnace. It is to serve to replace the ordinary blast furnace. 

From early times efforts have been made in countries rich 
in ore and water-power, but poor in fuel, to replace the fuel used 
in the blast furnace for the production of heat, by electricity, 
and so lower the fuel consumption. In the electrical process of 
pig-iron production there only remains about one-third of the 
fuel consumption necessary in the ordinary blast furnace, and 
this is for reduction only. In this way about two-thirds of the 
fuel cost is saved. At the same time the large blowing engines 
of the ordinary blast furnace are not required. These are the 
two important things that promise success to a good solution 
of the question of the electrical smelting of iron ore. 

Even in the introductory period of practically useful electric 
furnaces we find that they were first adapted to the production 
of pig iron. The Stassano furnace is an example which was 
originally only constructed for the smelting of ore. It is shown 
in Fig. 44, which clearly brings out how similar it is in construc- 
tion to the ordinary blast furnace. Stassano's experiment was 
unsuccessful, and we have seen how he turned to the method 
worked out in the meantime at La Praz by Heroult, for the 
utilization of scrap. Tests were also made in those parts of 
France having abundant wjater-power. Here Keller and Heroult 
were occupied with the question, and many reports and discus- 
sions of their experiments appeared in the journals in the middle 
of the last decade. 

The furnace used by Keller is shown in outline in Fig. 105. 
Two shafts are joined at the bottom by means of a channel. 
At the base of each shaft is a carbon electrode, these electrodes 


being connected by means of an outside cable. A carbon elec- 
trode is hung in each shaft. At the beginning of the operation 
the current flows from one carbon electrode and through the 
charge in the corresponding shaft to the bottom electrode. 
From here it goes through the outside cable to the bottom 
electrode of the other shaft, through the charge, and to the 
second electrode. As the smelting proceeds the connecting 

channel becomes filled with 
molten iron. As soon as a 
connection is made in this 
way between the two shafts 
the current flows through the 
molten material, which offers 
a much lower resistance than 
the two bottom electrodes 
and the outside cable. In the 
middle of the channel is the 
tapping hole. 

In a later construction 
Keller had a third small 
FIG. 105. carbon electrode, which was 

lowered into the connecting 

channel, and was used to keep the metal there thoroughly liquid. 
Extensive tests were made with this later furnace at Livet in 
1904, at the time of the visit of the Canadian Commission under 
Dr. Haanel. 

Of lesser importance were the tests carried out by Heroult, 
at La Praz, in the presence of the Commission. German Patent 
142,830, 1902, shows that Heroult had not left the subject of 
the smelting of ore in the electric furnace without attention, 
although he worked, at first, to develop a process for using scrap- 
iron and steel. This patent was granted on an electric furnace 
with electrodes built in the hearth and the shaft. It is shown 
in Figs. 1 06 and 107. It was not successful, and Heroult in his 
tests before the Commission mostly used a simple type similar to 
one-half of the Keller furnace. His average production with such 
a furnace at that time was 7.82 metric tons per 1000 E.H.P. days. 



In the time immediately following the visit of the Canadian 
Commission no further experiments in the line of pig-iron pro- 
duction were made in Europe that are worthy of notice. The 
general attention was devoted to the production of electric steel 
and iron for the very good reason that the results so far obtained 

FIG. 106. 

FIG. 107. 

in the production of pig iron showed no promise of success in 
Europe, for a great number of years, in competition with the 
highly developed ordinary blast-furnace process. 

On the other hand tests were continued in Canada, to which 
country Heroult went in December, 1905, his experiments being 
made in January, 1906. They were mostly carried out with the 
furnace shown in Fig. 108, consisting of a crucible with a shaft 
above it. The bottom, being made of electrode carbon, forms 
one pole, the other being a hanging carbon electrode. This 
electrode had a length of about 5' 10.8", and a cross-section of 
about 1 6" x 1 6". The maximum current was about 5000 
amperes, with a pressure of 35 to 40 volts and a power factor of 
cos = 0.9. The results under normal conditions were the 
production of about 11.5 metric tons of pig iron per 1000 E.H.P. 
days. Although good results were obtained with various ores, 
judging from a metallurgical standpoint, it was seen that an 
electrode entering the furnace with the charge would not satis- 
factorily solve the problem. For instance the electrode frequent- 


ly rose higher and higher in the shaft of the furnace, so that the 
material in the bottom got colder and colder. This was caused 
by the charge becoming too dense, and not allowing the gases 

to escape easily enough. 
In this way the resist- 
ance between the two 
poles was lessened, and 
the voltage remaining the 
same, the upper electrode 
rose in the furnace. It 
also brought about con- 
siderably higher electrode 
consumption. To sum 
up the question, a suc- 
cessful electric shaft fur- 
nace was not solved by 
the experiments made in 

In the spring of 1907,- 
experiments with electric 
pig-iron production were 
begun in Sweden. Messrrs. 
Gronwall, Lindblad and 
Stalhane together formed 
the "Electrometal" 
Company, with the aim 
of building and selling 
electric furnaces. The 
tests which will now be 
considered in detail were 
carried out by them at 
FIG. 108. Domnarfvet. According 

to Yngstrom's careful re- 
port in the Jern-Kontorets Annaler, No. 9, 1909, a current of 
7000 volts at 60 periods was used. With this current a 900 
HP motor was driven, directly coupled to a 25 period, three phase 
generator. From this generator the current was led directly 



to the transformers which were arranged near the furnaces and 
served them. Here also a switchboard with the necessary 
measuring instruments was set up. These included a watt- 
meter, three ammeters, and one voltmeter. Underneath were 
the hand wheels for regulating the electrodes. 

Gronwall, Lindblad and Stalhane first made use of the results 
of the former experiments. They therefore endeavored to 
completely obviate the use 
of hanging electrodes, and 
to keep the current in the 
hearth of the furnace. Fig. 
109 shows the first test 
furnace, which was built 
to take single phase cur- 
rent. Each pole consists 
of a copper plate carrying 
a graphite block. These 
blocks are hollowed and lie 
outside of the furnace 
proper. Channels lead from 
them into the furnace 
which, when filled with 
molten iron, serve to con- 
duct the current to and 
through the charge. Be- 
sides these two conduction FIG. 109. 
channels that are arranged 

on one side of the furnace, there is a third one, as may be seen 
in the illustration, and which serves for tapping the furnace. 

After charging, the furnace is run precisely as an ordinary 
blast furnace, until a considerable amount of metal has collected 
in the hearth. This insures good conduction from the carbon 
electrodes to the interior. The blast is then stopped, the current 
switched on, and the electric heating begun. 

The course of the current was arranged as follows: It 
entered at one pole and passed through the metal lying over it 
into the metal in the channel at one side of the furnace proper, 


from here through the charge to the metal in the channel at 
the other side, and so to the outgoing pole. Heating is brought 
about through an overheating of the liquid contents of the 
furnace on the one hand, and the resistance offered by the charge 
on the other. This should furnish heat sufficient to smelt the 
ore. The hearth was made of quartz. In operation it only 
lasted a very short time, so that the furnace could not be operated 
for very long. This was because its wave-like surface offered 
conditions favorable to attack, and brought about quick destruc- 
tion at the high temperatures reached. 

The first necessity was to rebuild the furnace. This was 
done in such a way that the electrodes led into the furnace from 
opposite sides, as is shown in Fig. no. At the 
same time it was hoped that the use of mag- 
nesite in the hearth would give better service. 
This, however, was not the case, for the reason 
that the magnesite, a fairly good conductor 
even at ordinary temperatures, became too good 
a conductor at a high temperature, and the ex- 
periment had to be stopped. This second test 
1 showed the impossibility of satisfactorily lead- 

FIG. no. m g the strong current necessary for heating a 
shaft furnace into the charge from the bottom. 
This style of furnace was therefore rejected. 

The third test furnace approximates in form the one already 
proposed by Heroult in his patent of 1902. It is shown in 
Fig. in. The shaft-like construction is furnished with three 
electrodes, of which one forms the bottom, while the two others 
are arranged on opposite sides at a medium height. The direction 
of the current can be so arranged that it either flows horizontally 
from one shaft electrode to the other, or else goes out through 
the bottom electrode. In operation the shaft electrodes were 
destroyed so rapidly that they were replaced by ordinary water 
cooled electrodes with continuous feed. With this arrangement 
considerably better results were obtained, but the walls near 
the shaft electrodes were so rapidly destroyed, because of the 
intense heat generated, that this style of furnace was also rejected 



as unsatisfactory. It, however, pointed the way to a good 
solution of the question. If care was taken to keep the intense 
heat, which is produced where the electrodes and charge come 
in contact, away from the walls, then more favorable results 
and a greater furnace life would be obtained. These considera- 
tions led to surrounding the electrodes directly with the charge, 
so that the heat, which was formerly lost through the walls, could 

now be used for heating the charge, 
and at the same time a much 
greater durability of the furnace 
walls was obtained. 

The 1909 test furnace is shown 
in Fig. 112, the lower part of 
which may be considered as the 
final form of the electric shaft 
furnace. This is the furnace of 
Gronwall, Lindblad and Stalhane. 
It has three electrodes penetrat- 
ing the roof of this hearth and 
is in general very similar to the 
ordinary blast furnace, except that the tuyeres are replaced by 
electrodes. The results show that this construction in its 1911 
and 1912 improved form is the most complete and suitable 
produced, and is the only one worthy of serious consideration. 
A detailed description is given below. 

The smelting part of the 1909 furnace forms a large crucible 
or hearth 7' 4^" m diameter, 4' n" high. It is lined with mag- 
nesite. The shaft of the furnace is arranged above the hearth, 
and has a height of 17' with an interior diameter of 4' 3" at the 
widest part. The shaft is supported by a steel framework 
resting on six iron columns. This makes it possible to independ- 
ently repair the hearth. The charge falls from the shaft into 
the hearth through an opening arranged in the roof. It forms 
an angle or slope of about 50 to 55. This produces a free space 
between the charge and the roof and walls of the hearth, on 
which the greatest importance is to be placed It serves to 
cool the electrodes and the walls of the furnace. To help in this 

FIG. in. 


purpose the cool waste gases from the top of the shaft are taken 
and blown through tuyeres into this cooling space. This method 
also brings heat back to the furnace, and so gives a better heat 
efficiency. As Fig. 112 shows, three carbon electrodes penetrate 
the roof of the hearth. In the 1910 model four electrodes, and in 
the 1911-1912 model six electrodes, are used. The early electrode 

FIG. 112. 

consisted of two carbon blocks 13" square, so that the total cross- 
section is 169 sq. ins. The electrodes are made in Sweden from 
retort carbon, and permit the use of a current of 25.8 amperes 
per sq. in. 

The electrode holders consist of strong steel frames. These 


are provided with several wedges by means of which the copper 
plates that carry the current from the cables to the electrodes 
are firmly pressed against the latter. The electrodes are operated 
by hand, and the part projecting from the furnace is provided 
with an asbestos cover to prevent oxidation. The openings for 
the electrodes have water-cooled seats, and arrangements are 
provided to prevent the escape of gas. 

When the furnace is put in operation it is run, at first, exactly 
like an ordinary blast furnace. The electrical heating is only 
used later. The furnace now described was run, with slight 
interruptions, from May 7, 1909, to the end of July. The follow- 
ing notes, taken from the account of the operations, are of special 
interest. At the beginning of the electric heating the current 
goes chiefly through the upper part of the charge, which means 
that the largest amount of heat is produced immediately under 
the roof, which is strongly heated and partly destroyed. One 
reason for this is that the lower part of the charge is colder, and 
therefore offers greater resistance than the upper part. The 
conditions were greatly improved as soon as the waste gases 
were blown in. The temperature of the roof was lowered, and 
the hottest zone sank lower and lower. The result was a lower- 
ing in the resistance of this part of the charge, so that the current 
found a more favorable path, and was concentrated in the lower 
part of the hearth. When this condition was once reached a 
five days' interruption of the gas-cooling brought about no change 
from normal running. 

During the operation of the furnace no big fluctuations of 
the current were noticed/ and even during tapping the instru- 
ments remained steady. This leads to the, conclusion that the 
resistance of the charge was very constant. The electrodes 
required very little attention. They were regulated once a day 
on the average, and in one case they were not touched for five 

The maximum current amounted to 9000 amperes per phase. 
With 25 cycles a power factor of 0.8 to 0.9 was obtained, with 
60 cycles of about 0.7, and other calculations gave 0.535. Natu- 
rally with a fixed cross-section of electrodes the amount of energy 


that can be used is dependent upon the permitted potential, 
which in its turn depends upon the resistance of the charge. 
The higher this resistance, the higher can the voltage be without 
the strength of current overstepping the permitted maximum. 
It is therefore of interest to know how to influence the internal 
resistance, and this consists in the choice of the proper amounts 
of ore and fuel in the charge. In the following table are given 
the strengths of current reached with various burdens, and with 
fixed voltages. 

Charge with 

between two 
Two Phases 
in Volts 

per Phase 

with an 
Average Cos 
</> = 0.85 
A = 1.73 ei cos0 

Coke in excess 
Coke not in excess 



480 kw. 
46s " 

Too little charcoal 



sss " 

Sufficient charcoal 



603 " 

Too much charcoal 



S36 " 

Too much coke and charcoal 



471 " 

Sufficient coke and charcoal 



536 " 

The operation of the furnace was very simple and uniform, 
the metal being tapped about every six hours. When judging 
the efficiency of the furnace it should be remembered that the 
following sources of loss are to be considered: 

1. Cooling of the electrodes with water. 

2. The ohmic resistance of the conductors and contacts. 

3. The radiation from the furnace. 

The total loss amounted to from 230 to 270 kw., the higher 
value coming at the end of the run. The loss is divided about 
as follows: The water cooling carries away from 118 to 225 
kw., which, with a power of about 500 kw., corresponds to a 
loss of about 25 to 30%. Overcoming the contact resistance 
takes about 40 kw., and from no to 180 kw. are lost by 
radiation. The electrodes lose 5.8 kg. (12.8 Ib.) per metric 
ton by burning away, the total consumption being 13.8 kg. 
(30.4 Ib.) per metric ton. From another source (E. F. Ljung- 



berg, Metallurgie, November, 1909), the consumption of elec- 
trodes through burning is 8.8 kg. (19.4 lb.), and through waste 
ends 13.9 kg. (30.6 lb.), a total of 22.7 kg. (50 Ib.) per metric 
ton. This large difference between the loss by burning and the 
total consumption is brought about by the electrodes not being 
completely burnt, and the ends having to be replaced by new 
ones. There is no loss from stub ends in the later designed 
electrodes which are screwed together. 

The maintenance cost of the furnace could not be determined 
exactly, but the furnace worked satisfactorily for 85 days without 
a stop. The weakest place is the roof of the hearth, which is 
exposed to the intense heat generated at the electrodes. Accord- 
ing to Ljungberg, 891,623 kw. hours were used to produce 280 met- 
ric tons. This means 0.492 h.p. years or 3184 kw. hours per metric 
ton of pig iron. This is a high figure and has since been lowered 
to less than 2000 kw. hours, or 0.31 h.p. years, on long runs.* 

The following tables give the efficiency obtained during the 
test with different burdens: 

Carbon Consumption 

Amount of 

Real Power 

Charge No. 

(pure carbon) 

Energy Neces- 
sary with the 
Given Carbon 







kw. hrs. 

kw. hrs 

365 days 


2 5 2 



























The economy of making pig iron in the Gronwall, Lindblad 
& Stalhane furnace is given in the chapter on operating costs, 
hence the following table by Catani is of interest. It is quoted 
from Neumann, Stahl und Risen, 1909, p. 276. This table 
shows how high the price of current per h.p. year can go, with 
coke at a fixed price, for the electric shaft furnace to compete 
favorably with the ordinary blast furnace: 

* See page 238. 


Pig Iron Produced 
per 24 hrs. per H. P. 

Coke Price 












Price of 






12.19 1 
15-23 1 







h.p. year 

For comparison with the foregoing figures of the first tests 
in 1909, the 1910-1911 tests results are here recorded. 

The November, i9io-April, 1911, test furnace of 2500 h.p., 
of Gronwall, Lindblad & Stalhane, is shown in vertical cross- 
section by Fig. 124. During the run the furnace was operated 
with four electrodes penetrating the roof, the furnace being 
operated with two phase current, from a three phase circuit by 
means of Scott connected transformers. The incoming current is 
10,000 volts, three phase, 25 cycles. The secondary volts can 
be regulated between 50 and 90 volts from the high tension side. 
The arrangements are such that the different phases can work 
simultaneously with different voltages. The method has greatly 
facilitated the working. Regulation is also had by different 
switching from the low tension side. The newer 3500 h.p. 
furnaces for Hardanger, Norway, using coke instead of charcoal, 
have the following dimensions: 

Diameter of hearth 3 meters = 10 ft. 

at ring 1.5 = 5 " 

" at boshes 2.15 " = 7 " 

Height of shaft 12.0 " =40" 

Total height of furnace 13.7 " =45" 

These Norway furnaces are somewhat different from the 
Trollhattan furnace. The volume of the shaft is smaller, but 
its diameter is greater than the corresponding shaft of a charcoal 
furnace. The coke in the charge gives it greater conductivity, 
so that a lower voltage is used. 

The ratio of volume of charge per day to shaft volume has 
been taken at 1.55, and the furnace volume has hence been 


made 38 cubic metres (about 500 cu. ft.). The furnace hearth 
is lined with magnesite. The general contour of the furnace 
walls and roof over the hearth can best be seen by consulting 
Fig. 124. The roof is cooled as described under operating costs. 
The gas that is blown through the tuyeres is cleaned in a water 
scrubber in the latest designs, as shown in Figs. 126 and 127. 
The electrodes used during the beginning of 1911 were built 
up of 4 carbons 2 metres (6^2 ft.) long and 330 x 330 mm. 
(13" x 13") section arranged to form an electrode 660 x 660 
mm. (675 sq. in.) section. 17,000 amps, is the permissible 
maximum or 25 amps, per sq. in. Toward the end of the year 
this has been changed to a cylindrical electrode of 600 mm. 
(23.6 in.) diameter, which is gripped much shorter than formerly 
(see Fig. 125), thus saving 40 kw. The square electrodes were 
supplied by both the Plania Works of Ratibor, Germany, and 
from the Hoganas Works, Sweden. The 600 mm. round elec- 
trodes have lately been furnished by the former works and by 
Siemens Bros. & Co., Litchenberg, near Berlin. The upper part 
of the electrodes is covered with sheet asbestos and thin sheet- 
iron, and the top surface is covered with a thick layer of ground 
asbestos and silicate of potash. They also have a water-jacket, 
beneath which gas was blown to cool the roof (see Fig. 124). 
This practise was not long continued, as the CO 2 burned holes in 
the electrodes. 

When starting the furnace, it is thoroughly dried out with 
wood and charcoal fires, and heated up electrically by filling 
the hearth with coke and turning on the current. About 3 
weeks is taken to burn through an electrode above the so-called 
" stock line." During January, 1911, the average voltage on 
each phase was 62.6 volts, and the average current per phase 
14,449 amps. The average reading on the wattmeter was 1535 
kw. ; the power factor was consequently .88 + % The efficiency 
of the furnace has been greatly increased since the tests were 
made with the 800 h.p. furnace at Domnarfvet, and is discussed 
under operating costs. 

The following table indicates the efficiency obtained during 
the tests as indicated: 



1 |S. I'S - S, o S 



2 Z 


oq 04 10 oq o^gaJoT^^^T 

IO IO rj- CO *Oi-J'5 l O|-4'5lO 

a t 

3* * 


O ' X ^ r< <U ^5 



O4 OO TJ- M vO iO w 
OO O cOcOO O O-< 




vO ON co co >"* vO ON 
ON ^t" O4 '^ ON co ^ 

O4 O4 04 O4 04 1-1 HH 


r ^ 


!'!!! ! 




iO ON O oo M oo 04 

1-1 GO VO O4 O VO tN. 





1 * c h 

O rj- O 04 co O 
O 04 04 10 


ON -^t- ON vO 00 00 
of CO 

vO O CO O 04 HH oo 
_ co 00 co 04 Tt- rt- 
Q\ 00 ON ON ON t>. t>. 




t^. co l^ 04 O ON 
iO vO ^O vO OO vO ON 
1-1 t>. TJ- O4 M CO CO 
rj- CO ^t- rf rj- co co 




IO ts, IO T|" 

O CO 04* vo' 04 



o ^ 

CO t>- CO iO M 
ON r^ 04 O4 O 
04 04 CO CO CO 





1 J 

CO O 04 Ol 00 
HH M CO t^- 04 



c ti 

l^ ON CO M ON 

HH vd ON co 04 
^* co ^" ^" ^* 


HI CO 00 04 ON 


,! J 

in ^J- o oo t^ oo 10 

O 04 00 iO O4 VO VO 
01 04 t^ Tf CO M M 


H " 




O vO 04 vO O 04 
i-* iO ^^ O O O 
C4 04 04' CO t^- O 

w 2 




l^ vO O 04 -3- O 
iO O "p ON iO vo 
i/^ 10 ON r*^ ^ t^ 

vo vO ^" ^^ vO vO 


ffli frj W W < S H S 2 I I 2 F 


Comparing the last two sets of figures with the first four sets, it 
will be seen what a great improvement has been made during 
1911. Comparing the above with the 1909 tests shown on page 
235, the improvement deserves the recognition it has received, 
in that over 30,000 h.p. of these furnaces have since been built 
or are building. 

It is interesting to know exactly what the first large 2500 
h.p. furnace installation of Gronwall, Lindblad & Stalhane cost 
at Trollhattan, which has a daily capacity of about 20 tons. The 
furnace house is of steel construction, and brick and both furnace 
and electric equipment cost more than a subsequent similar 
installation would, as this was the first one of this size. The 
cost was as follows: 

Excavation, railway connection, water-pipes, scale, etc. $10,727 

Buildings: Furnace house 14,735 

Charcoal storage-house 6,032 

Crusher-house, office, laboratory, shops 3,96 1 

Furnace 13, 1 1 1 

Electric equipment 13,782 

Cable and wires 3,832 

Gas-motor, pumps, reservoir 3,222 

Crushers 1,011 

Transformers 3,433 

Motors for crushers, etc 1,724 

Laboratory equipment, furniture, etc 10,430 


In order to give an idea as to the size of the necessary' plant, 
it may be said that an output of 10.65 kg. (23.45 Ibs.), per h.p. 
day corresponds to a power consumption of 1736 kw. hrs. per 
metric ton. With a daily output of 300 metric tons this would 
need about 35,000 h.p. at the furnace, or -about 38,500 h.p. at 
the power station, when allowing for a long transmission line. 
If a plant is built for $50.00 per h.p., it would require a capital 
of $1,925,000. Allowing 9% for interest and amortization, and 
3% for taxes, etc., each h.p. year would cost about $6.00 at the 
power station, or about $7.50 at the furnace. 

A complete furnace installation for 300 tons would cost 
about $500,000 and consist of six furnaces of 7,000 h.p. each, 


one furnace remaining in reserve. This estimate is based on 
the installation costs already obtained, but each furnace would 
be larger. 

It only remains to mention that because of the very favorable 
results obtained at Domnarfvei, 1909, the Jernkontoret of 
Stockholm has acquired the patents of Gronwall, Lindblad & 
Stalhane. The British owners of these patents are the Electro 
Metals, London, whose American and Canadian representative 
is Dr. D. A. Lyon, Pittsburg, Pennsylvania. 

The following iron ore reduction or electric pig-iron furnaces 
are built or building: 





O 0) 4) O 

CJ CJ *>| K/j u 

O <U 0) 0) 



rrt rrt O rrt 

J3 J3 O U jS 

h o o o 
J3 U U U 







10 10 o o o 

8 8 8 8 E 

! 8 8 









|f|f jf|f|i 

Jf JtJtJ! 

1 1 




> 2 M 

10 10 10 1 

! | 

Q "" 


^a ^ JD * ^3 w x 

i 03 

: -I 

D _ fl 
H Q O 

3 3 p ^ 3 w 

^ o 

3 10 

D O " 

o "^ 

,) CM j2 CO js CM J3^ 



o3 03 rt o3 



1 1 

ill ill iL 

O ^O JO 

U l-i 

"of 15" "3 

K* K*" 

C C G 

C o3 o3 

<L> 0) <U 

OJ o3 oJ 

^r* *<* <<* 

"S CO ' C/2 

<d *^ ^ 


^ "^ *+* 


^ o o 

C G C 
bfl W) U) 



a a 



-J-J *~~> *~~i 


o r 

'^ C3 D <i> 

r-^ -4 ^-4 

CJ o 

-"-I 0) ^4 ^ 

u^ 9o o^ 

^^ rj 

o o o 

J rt S* 

H s 5 5 S 


co "1 g 

-M ^5i Ji} *O 

2i 4-r u U 

T en | 

2 > ^ - c^ 
1 1 1 x 1 >, 

CD J J ' J2 C 

<y D o rt 

w cn . cn . en . T 

, -s -s g -s g 1 



^ O ctf ^* cd ^ crj 

^ Q K K ffi 

E < < < en 

1 1 


w CS CO rf 

<O t^ 00 O^ JH 

p ^ 

* c^ 


>> ^j 


^-o g 

|sl S 

d> ^ 

w ^ 

H-J C 


l ^ 

C *} *^ C/5 



K *PQ 



IN addition to the methods of furnace construction previously 
described there are naturally a tremendous number of proposals 
for the design of electric furnaces. This is best brought out by 
the many patents that have been issued both for arc and induction 
furnaces. Although the literature of such patent papers may be 
very entertaining, and is indeed very often instructive, yet a 
consideration of the many proposals does not lie within the 
scope of this book. 

Most of them are only proposals and will never be tried out. 
A smaller number disappear quickly after a trial and leave no 
trace, while the third and smallest part stand trial in one or 
another plant. They make possible the saving of the license 
fee for a successful furnace, but most of them cost enormous 
sums for experiments, and very often complications develop 
when they are put in operation. 

Although they are not for the most part of value to many 
people they yet have the advantage that they help to spread the 
knowledge concerning the properties of electric furnaces further 
and further. On the other hand, it is naturally only through a 
fresh consideration of the new methods of construction that a 
further perfecting of the old or even new ways can be found for 
reaching the wished-for goal. 

On this account, therefore, it is perhaps justifiable to con- 
sider at least a few of the furnaces differing in construction from 
those used most frequently today. Another reason is that 
one or the other of them are sometimes discussed in the technical 

Under the heading of arc furnaces comes first that of Chapelet, 
which is in use at the plant at Allevard (Isere). It is shown in 

Fig. 113. We see that the current flows in an arc to the bath 




from a hanging regulated carbon, similarly to the Girod furnace. 
From the bath it goes through a horizontal channel to a hanging 
cast-iron electrode that touches the channel. This constitutes 
the peculiarity of the furnace. It is not apparent that this 
arrangement offers any advantage over that of the Girod furnace. 
In the first place the furnace construction is much more difficult 
and not so accessible as that of the Girod. Further it is to be 
feared that the metal in the 
channel between the outer 
electrode and the bath will 
force up the furnace bottom, 
except that part which is 
not molten, because of the 
influence of the water cool- 
ing used for the iron elec- 
trode. This will bring about 
difficulties in maintaining 
the lining, since repairs to 
the horizontal channel are 
scarcely possible. The meth- 
od of working is exactly 
the same as that of the 
Girod furnace, that is to 
say, that heat is produced FlG - II 3- 

exclusively by the arc, the 

resistance offered to the current by the molten material not 
being of any noticeable value. 

The details of construction offer little that is worthy of 
attention. Water cooling is used at the opening in the furnace 
roof for the entrance of the carbon electrode, at the outer iron 
electrode, and also at the carbon electrode connections where 
the current passes from the copper cables. The cylindrical 
furnace roof is removable. The openings in the front part of 
the roof are used as working doors, as shown in the illustration. 
There are several of these furnaces in Allevard, but, according to 
Coussergues' report, only one is in operation. 

The Keller furnace, shown in Fig. 114, has still greater 


similarity to the Girod furnace. The only difference is a special 
arrangement of the bottom electrodes. While Girod, as we have 
seen, uses several water-cooled steel electrodes that are dis- 
tributed over the bottom surface of the hearth, Keller uses a 
furnace bottom formed of a so-called mixed conductor. As 

FIG. 114. 

seen in the illustration this bottom consists of a water-cooled 
iron plate over the whole surface of which are set a number of 
evenly distributed iron rods from one inch to 1.18" in diameter, 
between which magnesite is rammed. This is, in itself, a fairly 
good conductor. In this way a semi-refractory bottom is formed 



with a conduction between that of iron and magnesite. Accord- 
ing to Keller's results such a bottom is practically unmeltable. 
It is questionable whether his electrode arrangement offers any 
advantage over that of Girod. It depends on the durability of 
the furnace hearth in the two cases concerning which only work 
under practically the same conditions can give conclusions. The 
production of heat in the two furnaces is in no way influenced 
by the bottom electrodes. The uniform composition of the 

FIG. 115. 

whole furnace bottom in the case of the Keller furnace will not 
bring about the profitable circulation of the bath found in the 
Girod furnace. 

In this case also the original Girod is to be preferred to the 
newer Keller furnace, provided that the bottom will last as long 
in the first case as in the second. 

Often one finds in the patent papers the endeavor to increase 
the resistance of the bath by means of a suitable shape of hearth, 
and so bring about an additional resistance heating. As an 
example, the Nathusius furnace may be mentioned. Fig. 115 
gives a section of this furnace taken from the patent papers. It 
shows a number of electrodes of changeable polarity arranged 


above and below the melted material. In this way the current 
can be forced to flow through and around the molten bath. 
According to the description given with the drawing, the current 
flows first from the upper middle electrode b through the slag 
covering h and the upper layers of the metal bath to the upper 
outer electrodes a and c, second from the lower middle electrode 
e to the lower outer electrodes d and /. In addition, however, 
the current ought to travel from the outer upper electrodes a 
and c to the outer lower steel electrodes d and /, so that the bath 
will be enclosed by heat-producing currents. 

The whole arrangement, as is immediately apparent, repre- 
sents a combination of the Heroult and Girod furnaces. In the 

FIG. 116. 

first place it is presumed that it is possible to heat the bath by 
current led in through electrodes which have a much smaller 
section than that of the bath. This is naturally altogether 
impossible if the electrodes consist of carbon, as is the case with 
those arranged over the bath, which has an excessively high 
resistance in comparison with the fluid metal. In addition it 
can be shown that it is impossible to bring about much heating 
by means of the water-cooled electrodes, for their section in 
proportion to the bath is so small that the higher specific resist- 
ance of the bath can have no important influence. Fig. 116 
shows the practical arrangement of a Nathusius furnace that 
differs from the drawing in the patent papers because of a 
simpler and therefore better form of hearth. Here a direct 
heating of the bath, by means of the bottom electrodes, is not 



FIG. 117. 

taken into consideration because of the greatly increased section. 
The arrangement of the water-cooled electrodes in the latest 
furnaces differs from Fig. 115, and according to Neumann's 
report in Stahl und Eisen, 1910, they have a diameter of 8.66",. 
and are covered with a layer of 
dolomite 7.87" thick. With the 
passage of the current this layer 
gives off heat, and so much as is 
not carried away through the 
bottom electrodes enters the bath. 
For increasing this bottom heat- 
ing an additional 150 kw. trans- 
former is used for a 5-ton fur- 
nace. Currents of a maximum of 
6000 to 8000 amperes are used, 
that enter the bath from each 
carbon electrode, when a three 
phase no volt current is em- 
ployed amounting to about 2500 

amperes. The direct heating of a metal bath u.8" deep and 
about 78.74" diameter is altogether impossible with these 

currents. It is therefore also im- 
possible with the present arrange- 
ment of the furnace to use the 
bottom heating alone, although 
this is advanced as a special ad- 
vantage of the furnace in question. 
After all, the small advantage 
that the bottom heating may bring 
about must be looked upon as 
dearly purchased when it is con- 
sidered that the Nathusius furnace 

shows a much more complicated construction than the Heroult 
or Girod alone, and uses practically the same method of heat- 
ing. Moreover, it has more electrodes than the simpler older 
furnaces and therefore has greater heat losses. In addition 
six conductors are used for the current as compared with three 

FIG. 1 1 8. 


for the Heroult furnace. Apart from this the method of con- 
struction does not appear as good as that of either the Heroult 
or Girod furnaces. 

In the sphere of induction furnaces the constant endeavor 
appears to be the production of greater movement in the bath 
of metal. Most of the proposals show an ignorance of the 
principles of the induction furnace, for otherwise the designers 
would know that in these furnaces a completely satisfactory 
mixing of the whole molten material is produced by the electric 

FIG. 119. 

and magnetic conditions themselves. We can, therefore, leave 
out of consideration all the proposed furnaces that make use of 
inclined channels of small section through which the hotter 
material ought to rise, while the colder should descend. Such 
an arrangement proposed by Gin is shown in Fig. 117. 

The Schneider-Creusot induction furnace, of which a section 
is given in Fig. 118, is worthy of notice. This furnace, however, 
has not been improved since it was first designed. Like the 
Gin furnace mentioned above, and which appeared much later, 
it shows an induction channel with several hearth-like widenings. 


All such constructions of induction furnaces have the disadvan- 
tage that extremely high temperatures must be produced in the 
narrow channels if the material in the hearths is to be kept hot 
enough. This brings about a very energetic attack on the lining 
at these places, and as a result high maintenance costs and 
frequent delays in the working of the furnace with the Schneider- 
Creusot furnace refining is only carried out in the hearth A , and 
the remaining metal is kept free from slag. The use of the 
small hearth B is therefore not apparent. The arrangement 
of the furnace cannot be called simple. With this furnace also 
great value is laid on the increase of movement in the bath due 
to the great differences in section, and this appears reasonable. 
For obtaining this circulation the furnace is built on three 
columns, two of which allow a rise or fall in the furnace, so that 
during the operation the heating channels or pipes can be in- 
clined at a sharp angle. 

The furnace at the Creusot Works is arranged for a one- 
ton charge. 

Other types of induction furnaces endeavor to increase the 
resistance of the bath, and so bring about an improvement in 
the power factor. The proposal of Gronwall, which is shown in 
Fig. 119, may serve as an example. We see here the ordinary 
channel of the induction furnace, greatly elongated on one side. 
This arrangement naturally brings about a considerable increase 
in the resistance of the bath, but it has the disadvantage of 
causing very great radiation losses. Further, it is impossible, 
according to metallurgical practise, to maintain the division 
wall that is necessary between the two parallel parts of the 
hearth, because no refractory material is known that will resist 
an intense heating from both sides. Further, it may be men- 
tioned that such a furnace can only be used for the melting of 
pure materials, for work with slags cannot be carried out even 
to the small extent possible in the purely ring-shaped furnaces. 
This proposal, also, has not yet passed the experimental stage. 

Roberston, in the November, 1911, issue of the Metallurgical 
and Chemical Engineering, writes of the Gronwall two phase arc 
furnace. This furnace is the invention of Gronwall, Lindblad 


& Stalhane. Having originally worked with various types of 
induction furnaces without great success they decided to design 
an arc furnace. This furnace operates with two phase current, 
having two vertical carbons passing through the roof, each one 
to a phase. See Fig. i iga. The current arcs from the electrodes 
to the charge, passing through this and then through the basic 
lining at the centre of the hearth bottom, to the neutral return 
which is a carbon block fixed in the bottom of the furnace. The 

top of this bottom electrode 
comes level with the brick- 
work so that it does not 
project into or in any way 
weaken the basic lining. 
The hearth of the furnace 
therefore is not broken by 
any projections. The fur- 
nace has three doors, one at 
each end and one at the 
spout. Either hand or 
FlG j j 9a automatic regulation is pro- 

vided for the electrodes. 

This furnace is of the tilting variety, being mounted in curved 
rails. Heat regulation is obtained by varying the voltage of a 
special regulating transformer. The normal working voltage 
is 65. 

As each phase of a two phase circuit is connected to one of 
the vertical electrodes the arcs are independently formed, so that 
if one arc is broken the other remains. This insures steadier 
running than if both arcs were in series as in the Heroult furnace. 
The arrangement of two arcs operating in parallel with a ne tral 
return through the bottom produces a vertical as well as a 
horizontal circulation in the metal bath, slightly different from 
that in a Girod furnace. 

Naturally there have been many attempts to combine the 
various types, such as the induction and arc furnace. Fig. 120 
shows one, and is that of Hiorth. (Such proposals originated 
at the time when the causes for the failure of the channel-shaped 



induction furnaces for refining purposes were not clearly known, 
and it was thought that the slag temperature was not high 
enough. In the meantime the successful operation of the 
Rochling-Rodenhauser furnaces has shown the incorrectness of 
this reasoning.) In Fig. 120 we 
see the channel of an induction 
furnace broken by a division wall, 
which is bridged by a stirrup- 
shaped electrode. This electrode 
should only just touch in the slag, 
and bring it to a very high 
temperature. This shows a com- 
plete ignorance of the probabil- 
ities. The unmistakable result of 
the proposed method of working 
would be a complete freezing up 
of the metal in the channel on 
the opposite sides of the elec- 
trodes. It would be impossible 
to introduce sufficient current 

into the bath through the electrode, with which to produce 

heat enough, by overcoming 
the bath resistance, to keep 
the metal fluid. 

Even so, Hiorth says that 
he does not consider this 
furnace construction to be 
valueless, still we do not find 
that he has used this method 
in the single commercial fur- 
nace which he has constructed, 
which is shown in side eleva- 
tion by Fig. 121. This is a 

FIG. 121. purely induction type of fur- 

nace with the primary winding 

in flat spools similar to the arrangement already proposed by 
de Ferranti and later again by Prick j excepting that Hiorth coils 

FIG. 120. 


both legs of the magnet. The coils, according to a paper read by 
Dr. Jos. W. Richards before the American Electrochemical 
Society, in 1910, are uninsulated copper coils, hollow and the 
lower ones water-cooled. The construction of the ' furnace is 
such that it may be tilted independently of the magnet. 

So far, Richards continues, Hiorth uses his furnace only for 
melting the purest obtainable Swedish Dannemora pig iron and 
Dannemora Walloon iron. Yellowish- white blast-furnace slag 
was being used as a flux. The contents of the furnace being 5 
tons, 3 tons were poured at a time and 2 tons left in to start the 
next charge. The details of a heat run-off are then given which 
are here omitted. We quote further: 

(Assuming 300 calories necessary to melt i kg. of steel the 
thermal efficiency of this melting operation is 55% and the 
furnace radiation loss calculates out 180 kw. at this temperature. 
It was stated that it took about 170 kw. to keep the charge 
melted when the furnace was kept up to heat over night.) 

The power factor varied from .80 at the beginning of the 
run to .57% at the end when the metal in furnace was 5.77 tons 
and at casting temperature. Current used averaged 395 kw. 
for 6 hours or 790 kw. hours per ton of steel. As low as 700 
kw. hours has been reached in this 5 -ton furnace on cold mate- 
rials. This furnace operates at 12^ cycles, 400 to 500 kw. at 
250 volts single phase. 

Other proposals consist usually of combinations that in most 
cases would bring about great difficulties in operation, and which 
offer no advantages over the original furnaces. Here belong 
those which take an ordinary metallurgical furnace, such as an 
open hearth, and operate it at certain times by a stoppage of the 
gas, and the use of carbon electrodes. Also a combination of 
converter and electric furnace, for instance a small converter 
with an arc furnace built in. With these combinations the 
conditions of operation have not been considered carefully 
enough. For instance an open-hearth furnace in comparison 
with an electric furnace has such a high roof, and large working 
surface of bath, that the heat losses when using carbon electrodes, 
even if only for the desulphurizing period, would bring about 



much too high costs. In these cases it is therefore much better 
to transfer the charge from the open-hearth furnace, or the 
Bessemer, to a special electric furnace by means of a casting ladle, 
and to stand the unavoidable heat losses. In this way cheaper 
and better results will be obtained than with any of the proposals 
mentioned above, none of which has been really seriously tried 
out up to the present. 

This book would not seem complete if mention were not made 
of the Baily resistance furnace for heating bars, billets, and 

The first experimental work, leading to the successful develop- 
ment of the present type of resistance furnace for reheating steel, 


FIG. 1210. 

was done in 1906 at the plant of the Transue & Williams Com- 
pany, Alliance, Ohio, and although subsequent work at this 
plant showed that electric furnaces for heating bars was feasible, 
the management of the plant was not satisfied by the cost of 
operation, owing to their lack of economically .developing power. 
In the fall of 1911 Thaddeus Baily, of Alliance, Ohio, built at 
that place an experimental laboratory and demonstration plant 
of 200 kw. capacity, and after some months of careful work 
built a furnace that has all the requirements necessary for a 
commercial electric furnace for the purpose of heating steel for 

The latest type of this furnace for heating bars and small 
billets for forging operations as shown in Fig. 1210, may be 


^*mmm, \ 

briefly described as follows: In the bottom of the rectangular 
fire-brick enclosure D is located the resistance material A com- 
posed of crushed coke; the bottom of the furnace is protected 
by a lining of high-grade magnesite material C, which will not 
flux with carbon, even at the high temperature of the electric 
furnace; the current is conducted through the furnace walls by 
means of graphite electrodes B-B, making contact with the 
resistance material or core immediately inside the furnace 
chamber; the heat given off from the core A heats the bars E 
lying immediately above it and in the same chamber, and which 
are supported by a ledge at the back of the furnace, as shown 
in Fig. i2ib. As the metal and the resistance material are in 

the same chamber, trie slight for- 
mation of CO from the coke keeps 
the heating chamber under a re- 
ducing atmosphere, and there is 
no loss of metal by oxidation, aad 
hence no scale forms while tfie metal 
is in the furnace. No coal, gas, or 
oil fired furnace for heating bars 
or billets has this advantage. 

The furnace is controlled by 
means of voltage change, the usual 
range required being two to one, 

and is performed by means of a special regulating transformer 
and dial switch, an increase in voltage making a proportioned 
increase in current flow, the input in kilowatts increasing as 
the square of the voltage, within usual operating limits. 

This type of furnace is ideal from an electrical standpoint, 
as the operating voltages are moderate, seldom exceeding no 
volts, and the power factor even in 60 cycle circuits is extremely 
high (seldom less than 98 per cent.) on account of the load being 
practically non-inductive. 

The temperature range and control are all that could be 
desired, as any temperature may be obtained and maintained, 
that is, within the limit of the refractory materials used for the 
construction of the furnace. 

FIG. 1216. 


Temperatures of 1750 C. in the furnace chamber have been 
readily maintained for long periods without other trouble than 
that due to the roof and walls of the furnace, which in this case 
were of silica and softened at the temperature named. 

As the temperature is proportioned to the input of elec- 
tricity in kilowatts, the control of temperature through the 
current regulating device described above is extremely simple 
and accurate, and when the furnace is once heated until the 
walls have reached an equilibrium the loss of heat becomes 
constant and may be reckoned in a definite number of kws. 
per hour, for any given temperature. With a definite weight of 
steel, being heated per hour to a certain temperature, requir- 
ing a known number of kw. hours, the voltage may be adjusted 
to give this required kw. input, and in such cases the tempera- 
ture of the furnace cannot vary. 

With an increase or decrease of heating capacity of the fur- 
nace, the voltage and kw. input may be readily changed to 
maintain the desired temperature. 

Temperature of interior at charging 2615 F. (1435 C.) 

Temperature of interior at withdrawing 2600 F. (1426 C.) 

Voltage of furnace 5 2 

Amperage of furnace 1000 

Power factor 99 

Indicated kilowatts 52 

Amount of metal charged, being 8 bars { , , . . 

I* in. (3.8 cm.) square by 18 in. (45 cm.) long.. . \ * P Unds (4 '' 7 "*> 

Time in furnace 20 minutes 

Temperature of metal, when charged 60 F. (15 C.) 

Temperature of metal, when withdrawn 2360 F. (1239 C.) 

Kilowatt hours consumed in heating metal 17 

Pounds (kg.) of metal per kilowatt hour . 5 pounds (2.25 kg.) 

Kilowatt hours per gross ton of metal heated 440 

Capacity of furnace, per hour 276 pounds (125 kg.) 

A typical operating performance on a 60 kw. furnace heating 
slightly over 3 tons daily, or 135 kg. (300 Ib.) of steel per hour, 
is given above; the table, showing a rate of metal heated per 
unit of current consumption of 2.25 kg. (5 Ib.) per kw. hour 
or 440 kw. hours per ton, with steel heated to 1300 C. and 


requiring 220 calories per kg. (400 B. T. U. per pound). These 
figures show a thermal efficiency of 67 per cent. 

The following list shows the furnaces of the type described 
that are either in operation or building: 

Names of Plants. 


Number of 1 
Furnaces. || 


Capacity per 

Canton Drop Forge Co. . 

Canton, Ohio 


60 kw. 

^OO Ib.I^S kgf. 

J. H. Williams Co 

Brooklyn, N. Y . . 


so " 

2OO " 90 " 

McKinnon Dash Co. . . . 
Transue & Williams Co. 

St. Catherines, Ont., Can. 
Alliance Ohio 



40 " 
60 " 

150 " 70 " 

j OO ' ' I T> S ' ' 

Electric Furnace Co. . . . 


40 " 

ISO " 7O " 

The Baily furnace is manufactured and sold by The Electric 
Furnace Co. of Alliance, Ohio. 

As a conclusion to this review, which is believed to embrace 
the most valuable proposals in the different spheres, it may be 
established that, until the invention of further types of con- 
struction, we have only to deal with those described in detail in 
the special chapters. These furnaces still show many weaknesses 
in comparison with the ideal furnace, yet they show that in those 
with the greatest simplicity the ideal has been closely approached. 



THE purely technical side of the application of electric fur- 
naces to the iron and steel industry has been considered in the 
foregoing chapters, so that now something may be said with 
regard to the economical questions of electric heating.* 

We have seen already in Chapter I that the development of 
electric furnaces is closely connected with that of electro- tech- 
nology. This is still the case when the question as to whether 
the installation of an electric furnace under certain conditions 
will be an economic success or not is under discussion. Then, 
indeed, the cost of the electric current, which is the heating 
agent of the electric furnace, is of real influence for the success 
of an electric steel plant. It must be taken into consideration 
that electricity, in by far the most cases, is much more expensive 
than the ordinary methods of heating, nevertheless this disad- 
vantage is more than equalized by other advantages. 

In this connection we may quote from Borchers' address 
before the Verein Deutscher Eisenhuttenleute, in 1905. "If we 
reckon the kilogram of carbon in coke at a high price, say about 
o.7i4c., then 1000 kg. calories will cost o.o88c. Very cheap 
electric power, namely at $9.52 per h.p. year, gives 1000 kg. 
calories from o.i67C. to o.2i4C. according to the number of work- 
ing days. This disadvantage of electric heat production is bal- 
anced by this condition; that the material to be heated, which in 
this case is the charge itself, accomplishes partly or altogether the 
transformation of the electric energy into heat. In a certain 
way it forms, of itself, the source of heat, while in all combustion 
furnaces the heat goes first to a mixture of gases, and from this 
to the material to be smelted." 

* For a more detailed discussion see Part II under "Costs of Operation." 



In the above example the price of power taken is very cheap, 
for with $9.52 per h.p. year, and assuming 300 working days in 
the year, the kw. hour only costs 0.1780. Such a low figure 
is only to be reached with the use of very suitable water-powers, 
while it is unattainable by using blast-furnace gas, provided that 
the blast-furnace gas is reckoned at a cost corresponding to its 
heating value. If this is done then it will usually happen that, 
even with the use of large gas-engines, the kw. hour cannot be 
furnished lower than 0.3570. to 0.7140. Still more unfavorable 
are the results if steam is used, although here, also, progressive 
engineering has brought about a constant cheapening in the 
price of current. For instance, in well-conducted central stations, 
with the use of large steam turbines, it has been found possible 
to produce the kw. hour at about 0.7140., when the coal does 
not cost more than o.4ic. per kw. hour. This is, of course, 
provided that the demand for power is very uniform, and free 
from variation, for otherwise the price per kw. hour is increased 
considerably. In this connection von Rizzo, in the Electro- 
technische Zeitschrifi, p. 596, 1910, gives a figure of 1.310., 
the power being produced by steam, and being used for operating 
a railroad with a very variable load. 

The prices given have reference, almost always, to large 
central stations, such as large iron and steel plants, city stations, 
etc. With smaller producing plants the price of current naturally 
rises considerably. It is therefore recommended that small 
plants should almost always be connected to some large central 
station for their electric furnace power, if the opportunity is 
there. Such stations today often furnish power for 0.9420. to 
1.4280. per kw. hour, which is a price that cannot be realized 
in small power stations, except with high pressure internal 
combustion oil engines. 

We see then that the source of power used for the production 
of electricity can affect the price of current, and therefore the 
production costs of electric steel. Also the way the current is 
used plays a very important part, and this depends in the first 
instance on the method of working. The following table shows 
how this method of working influences the power consumption: 


It requires for the production of: 

Pig iron, direct from ore 2,000 Kw. Hrs. 

Steel, direct from ore 3,ooo 

Steel from cold pig iron 1,500 

Steel from fluid pig iron 1,000-1,200 

Steel from cold pig iron and cold scrap. . . . 900-1,300 
Steel from molten pig iron and cold scrap. . 600-1,000 

Steel from cold scrap 600-900 

Refining of molten low carbon steel to make 

special quality steel (with very complete 

chemical purification) crucible steel 

quality 200-300 " 

Refining of molten low carbon steel to ordinary 

electric steel (electric rails) 120 

Retaining pig iron molten for foundry purposes 

(heated mixer) 50 " 

These values can naturally only serve as rough estimates, 
because the composition of the charge and the finished material 
are absolutely necessary for more exact figures. Further, more or 
less power will be used according to the efficiency of one or the 
other furnace, so that with the same charge and finished product 
different power-consumption figures will be given by two furnaces 
of different types. 

The wide limits given for working mixtures of pig iron and 
scrap are necessary because the power consumption is greatly 
dependent on the percentage of pig iron and of scrap used, more 
being necessary with an increase of pig iron. Further details 
on these points are given in the second part of the book. 

From what has been said it is apparent that the price of 
current becomes more of a determining factor (for the efficiency 
or non-efficiency of electric furnace operation), according to how 
many of the metallurgical processes necessary for changing ore 
to steel are carried out in the electric furnace. For instance, 
when fluid metal from a converter or open-hearth furnace is 
worked the cost for power, with an average unit price (o.476c. 
to o.7i4C. per kw. hour), is about 3% of the production cost; 
but it increases to 12% with the same price per unit, when scrap 
is worked. 

Finally we must remember that all of the furnaces in use 


today have certain special advantages. Unfortunately each 
type of furnace has also certain disadvantages. These disad- 
vantages are so closely connected with the methods of heating, 
that they must be allowed for. If electric steel production is 
entered into today one of the existing furnaces must be chosen, 
and its advantages and disadvantages purchased together. It 
is therefore not without interest to see how widely distributed 
the various types have become up to date. The following 
statistical tables date up to 1913. 

The tables show that the more important furnace types 
have already become so wide-spread that they must be considered 
to have passed the experimental stage. At the same time the 
electric furnace has shown that it is of considerable economic 
importance because it has enabled the production of the 
very best finished steel from low priced material. Until now the 
purest and therefore the dearest raw materials were necessary 
for this purpose. The tables clearly show that this great econo- 
mic advantage of the electric furnace is becoming known more 
and more. 

When we realize that the Stassano, Heroult, and Kjellin 
furnaces were first brought out in 1900, and the Girod and 
Rochling-Rodenhauser in 1906 and 1907, the wide-spread dis- 
tribution of these furnaces takes on greater importance. How 
quickly this distribution increases is also shown by the table, for 
in addition to 131 furnaces row in operation, 14 are under con- 

It may be concluded by pointing out that the electric furnace 
is already firmly established in the iron and steel industry, that 
the present development of electric furnace plants has been 
very rapid, and that an important future is assured. 



3 | 


*c3 S *^S ^ 

C "aJ ^ 

C3 O 4J 


^j ^ 

*-* u 

<u U3 to *- 

*n .y ^ n 







1 g . 

1 51 1 

S 8 5 3 




<u n 

5 M 1 s 

1 6-g.g S 

.3 Ui -H >>> n 

-M & 03 -0 ^ 


. w C 


-M o3 o 




5 8 


38^ . 

"B S - 

*T^ *rt a) ^ *^ 

'cd ^ "^ "c3 

^ > *c3 

gj """" 




8 6 

o fi a O o 

**^ (1) rH **^ 

"O C w 

I I o | 

-a -a 

2 ffi a 

G O C 

S s 

6 K 2 

a s 



aJ *Q ;< O c3 

u u o c: u 

8^5 S 8 

0,2 3 8 

P 03 

o i; o 



































t/> *2 *~* 




u a ? 

3 tJ-a 

J 55 1 


> <u 





- 1 




i-i < p u. 


a . 

a. > 


5 u 


a go 

"3 <u 

0> ^2 

(f j C" 

-M . rh -M 

c3 ^H c3 

nj K! ^ *o* 

.Si C ^ bfl.Si 
u c *^ ^ u 

ismarckiitte, U 
Silesia, German 

israarckhiitte, U 
Silesia, German 

works, Burba 































(a) ARC FURNACES (Continued} 

Use of 

Seamless tubes, 

rails, ordinary 

Seamless tubes, 

rails, ordinary 


^ 2 S 

G ^ 5 

^ , "* 

-- cn o3 *C 



Dynamo, die 
steel and fine 
plate castings. 



















Cold material 

and liquid OH 

Cold material 

and liquid OH 
Hot metal. 

Cold material 
and liquid OH 





















































g 'a.S 


U W M 

M i 

w a G 





















"Deutscher Kaiser" 

Bruckhausen, Ger- 

"Deutscher Kaiser" 

Bruckhausen, Ger- 
"Deutscher Kaiser" 

Bruckhausen, Ger- 

Gebr. Bohler & Co., 
Kapfenberg, Austria. 

Kartnerische Eisen & 
schaft, Ferlach, Aus- 


"o , 

CH rt 

- "cn 
cn 3 

2 < 

CQ j-r 


CX oS 

Gg. Fischer, Schaff- 
hausen, Switzerland. 








*This furnace was des 














































p /3 


cd *C 











c . 

c . 

c __. 


o Q 



QJ i 


-a 6 










-O -G 

TT J3 

*T3 J! 









3 t! 

3 oj 

3 oj 

*3 "*"* 








CT 1 CU 
.-, -C 





































1 g 













































M 3.S 


U W c 

l |1 




















Societa Tubi Mannes- 
mann,Dalmine, Italy. 

Societa Tubi Mannes- 
mann,Dalmine, Italy 

Soc. Electrometallurg- 
ique, Frangaise, 

3 CU 


r/T CO 
CJ en ~ 

bX) CU CH 

Usines Metallurgiques 
duHainaut, Couillet, 




^ w 

< -a 

C cu 
"o ^ 



I==H bJO 


aj u 
J3 U3 

.y ^ 

Obuchowsche Stahl- 
werke, St. Petersb'g. 

Aktiebolaget Heroult 
Electriska Stal, 

Kortfors, Sweden. 
Halcomb & Co., Syra- 


































V CL) 

















3 3 

2 '5 

3 a 




^ K 



S 1 4 ^ 

(f> en ^ J 

o -3 --S o ^ | 

o /3 S * 

i "n 

! t 

i g 












^ c ^ cD 

O -M O 4-< ^ ^ 

'- Jr ' Js ^3 "a "O "a> 
~ S S '3 ft '3 ft 



.zf -^ .2^ -c cr en cr w 

_I J J J 



: o o 




cs" cT 






03 oj o3 oj 
J3 43 J3 43 



a a, a a, 


<L> 0> O <y 




"So "w 2 2 


C C J3 ^ 


c/5 c/5 H H 


3 5? 








. * . . . 

2 2 


* if 

U 33 

. "^ 







CS 10 ^ I 




d" of d" . .S C T3 



.y P 

o ' r 
.y r oj 


gj GjOHO !y a \<f5 c/ ' 

(/j-t-T (/)4-T ~*o^X- 

4Jk / 4.4->k> ^f^ ^n 

{/) SM t/3 (/3 35 .2 .** * * c 

Welland, Canada, 
lectro - Metals Lt 

Welland, Canada, 
o. Mexicanade Ace 

y Chimicos, Mex 
o. Mexicanade Ace 

X Cd DH 

^ U e 
-2 S 


ll c <f<d 



fe E S < W 







to Tf 10 VO 




cs cs cs cs cs 








73 3 3 ^ 





2 222 2 






ffi ffi ffi ffi ffi 













*o c/r 

"O aT *^ co" 'O co* 

G *^ 

G *Jy 

G -3 G G ^5 


S ffl . 

>> ^ CO 

>, " co' 

. ta , ^ S . ^ . 

o ^ 

^ , ^jQ 

^ -H ba 

"rt ^ 

U, .Si) J-, . S) VH . S) 

o> o 

L* c\J C 

rt .2 G 

rt .2 C ctf .^ G rt .^ G 


C CU 00 

C O '-M 

Hjj |,| 


C O '-M G *CJ "-M C O * -M 

-H O tn .rtCUco .^cuco 




O O O 





1 6 





c^3 cd ctf 
42 43 42 








12 * 2 




"o o "o 




U U U 





|a| 1 | 




co Ui co co j_i co 










O Qt |H ^ O p ( *-*-* O 





JJ CU ^ -M CU 45 -M 






" J *S -^ J3 'S -^ 




^ H H 

2 so 

co -5 c 

O " 


o u w> 




' 3 ' 

' 2 ' 

. o o o 

M I-H 

- cu <u 

3 .S 

^ 3 G 

| g 

^ p C 3 $ G 3 ^C 

cr 'Sn 

c^ '5fl 

.2* ^ .5f bQ .2" ^ 

00 * i^ 

co -! hir' 

a, 5 ^ 

<L P 

cu H ^^ cu H ^^ cu H ^^ 


O 13 
fe U O 

? o 


^ ^3 

4^ -o -^ -n -32 -r? 

feWg feWg feWo 

(U - 

O <y 'J^ 

CJ CU -^ 

CJ C^ * f 1 O OJ " H U O " H O 

2 ^p J5 

rt - v ^ ^ 

C T3 vnj ^"2 ^ t2 vnj *T2 CH r Tj vflj *7^ ^ 

| -o II 'i || -^ || 

<u < fS 

J-( CU *^ PH 




U U U 




rl- 10 vo 










b b b 



1 "3 

























b "i3 




13 "o 







.2 5 


03 Cfl 

- JS 

c3 c> 































T3 >- 


1 i 






1 & 









IS -s 

*O .*""{ 


*O _^Z 





O ^ 




"o u 





















880 | 































































t/3 'p ~ 



W M 

So c 

H * o'S 

s z 


^ UJ 



00 n 

C tn 





* o 

01 4-> 

o s 

M o 




S . 




Oehler & Co., Aarau, 

Soc. John Cockerill, 
Seraing, Belgium. 

A. Stotz, Stuttgart- 

Oberhausen, Germ. 

Stalwerke Becker, Kre- 
feld, Germany. 









werke, vSchoeller & 
Co., Terniz, Austria. 



i > . 

d '5 










































"S 3 




en 0> 

Si OJ 



"c3 "o3 






G In 

*-M "3 

.S ^ 








cn o 

03 *- 


oS *-> 


"o g, 

















































CU flj 

en 35 

"a ", 

















bO ?i 





L (-| 


L f-] 




co H 





co ?.S 


to ^ 


ug g 1 

3 o'a 







^ d OT 





i .s i 

o c 

~^ ^ 

U ^ 

6 03 

U fc 

s s" 

<L> a, 

o'S S 
U x 


C W ^ en 





o -Q 
S S 


5 o 

O ^2 


o 6 




S y b 



1 * ^ 


Bethlehem S 
So. Bethlel 
U. S. A. 

Rubery Owe 
Steel Work 

oJ o3 

o '-' 


*rt "ol 

O t ~ l 

H., Bonn. 

H., Bonn. 





2 s 




















g i 














Use of 

en M en M 

u w u : 



d d 

u u I 

en en 

2 2 
"o *o ! 
U U 


ooioooooioioio >o 

WCNCOOCNi-ci-iMC'<C'< 01 



en CJO en oO en en x en en en en en en 
o3Co3Go3 o3o3 o3 o3 03 o3 o3 o3 
J3 .C -Cl -C -C -C .5 ^ JS 43 X 

a c5 a 03 a. a a, a, a c, a c. a, 

i C ^j G j CUCL) O O O O CL) O 

o2 <y o ll) oo o o cu a o o 
"rt"rt ^tS H H H H H H 



en ^ d 
1 '~ 

o : i 

OH g> 

H o'e 

: : : 





* * 1 .-g.-M ?i 1 1 d d? 

.PL, 3 T3 T3 T3 ^ -3 ^ A , J. 
>-o3 1 - 1 ' 1 S b/Oen en en ^ .^ cu 5 OJ^; 
O O ^ *^ CJ O O (*s \^ GJ - ' eU 4-> 

rt e S _g ""I J J ,.- ^- ^ g C/) rt 

O > O > n ^ ^ '^ -^ -S g^ajg^aJ 
^o^^; S 25 S . Si ^acr^<uc 

r- T 1 j^ frj _^ _j _* O S O -* FT ^'y'' *-^ LT. ( ~3r K^ 

_ C __. C ^ HjH ^3 ,_pH ^3 ^H <-* KO ^*^ ^1| l>** 

S^rtlHrt'-d'-d'-'d'^^'^^o r H o rH 

O ii O -M t -^ ^ c CJ ,s en en u, ^ ' u. ^3 ' 

a o a -o .< S .t> *o _o3 -a _o3 *c ^ 3 3 -g i: ^ -g is jH 

O O PH "o [T< *u ^ * ""* PH * o _ . 4^ QJ t 1 *!^j 



d d d d d d o d d d d 

cccccccccc c 

cnxenenencncnenenen en 
encncnencnencnenenen en 

CA}C/2C/)C/iC/iC/lC/)t/)C/lC/} C/) 



^ ^ 

1" gl 


'1 if ^2 

"o " 

O* fe 

Er > *G 

W 1 W OQ 

g* "to to 


"c5 "3 .2 
T> jg fi 

g. M *- 

. ^ 

03 o *2 H3 


G *Z* O 

fl ' 






.G T3 

S 8 


_C T3 





i C 




Q qj O3 






Z-H ^ 

Z < ^ ^ 

















Single phase 
four electrodes 
groups of two 

Same current 
one electrode 
and conduct- 
ing hearth 

Same current 
one electrode 
and conduct- 
ing hearth 

Same current 
one electrode 
and conduct- 

ing hearth 


o " 


U M bo 
" S ^ ^ 

*" ^ 






6 G 

U o 



tti S o3 
-2^ o 

1*? s 


<" 3 

O J 

"o <u 


S u 2 

4J QJ [i, 

\ry^ ^_ 

"- fe 



fe ^; >> 





o ^ 

1 1 

o rH 

H TJ i 
o -^ js 

JJ H- 1 +-> 



I s 



'I s 

o> -a; 3 


o> M - 

0) <U 3 

I 1 : 




















'.- 2 



s 1 



i- - 

X to 

>, en 

>, tc 



<U ~ en 

oJ rt 

c 'o 

o3 o3 
C 'o 

'-5 a 


G "0 
T3 ft 








W 13 








*2 w 
















g. | 



















11 ^f 




05 G 
P '13 

a ca 

$ bfl 

03 C 


1 .s 

*5 -^ 
S* c 


(U (LI 

e a; -u ^ 

| 1 5J -S ^-S 

If ^ 


c 5 J-; 


o3 C C -~ 
CO O oj 

Rj C G - 

CD O rt 


CO ^ 



I o "" 

l H 

! y u M 

|| || 

I C 











y3 i G 

.D u 35 

J2 A i oT u 

1 1 o g o 

^ W Ic 

o tw - 

o w - 



fc ^ ^ - 

j ^ , C 

. s ^ 


^ <D 4^ 


1 ^ E 

2 1 [ 

H K 4_> (-C 

\ * & 2 

4-i *"* 

- ra S "^ n 

l>v I/I & 

c/j 2J 

~ r Ji & 

| S | 

I eg g S S 

c ^ x ' 

O cu -^ 
h "O X i 

H s] 

3 <L) T5 
3 J 1 





































G , y i 


U U. 

= 3 




! -g C > G "5 




11 .! 

J jn~ J5 ^of 45 Jff 


"5 & t 

i, CO _ 

! . 

J M rt "S rt to 


J-H C 


; ^ 

^ ^ 


s y 



oJ q 
bfl t 

ji *bio^ s bi 



03 c 
g 1 

2 T 

H I- 

1 a 

5 i 

3 *C 

i ' a w ^ J 

! -Ill 1 

o "c 

! c 

.2"^ w "o 

U L 

) L 

) J U 

| sl 

8 * 

5 " 


5 MD Q 
fO O 



4 H 

r^ ro 

a, ba g 

1 bfl a 


03 C c 

J.S ? 

.I* 1 J? 03 .S 3 


CU rt C 
<U G 4 

I rt "c 

j,| ,! "g,| 


*5i S3 i 

' ^ i 

a3 fe a; 

C+-> t 


" E- 

; "3 E- 

o3 t^ o3 t" 1 o3 

o M 




; o be 

1 38 -S 

H g o C 

UK ^| 



* C OT 



5 Of 

i O 




vO n C 

OJ ^ 

.S i 



b < 

U ^ ,5 






5 - 
S &U 


E ^ fe CA 






S 03 ^ '"53 -2 

2 ^ r E 




en <U 


yX ^- C .-^ ^ 

* c 

O ^ 

-S x-H^ 


a ""O bic '' 



. 3 03 - 

03 > > 

C/l < 



< S c/i ' S <3 






o ^-o I 

' t-3 " M i* " M ''T1 


1-1 ^ > c 

u i_5>o>-<h-4feO u '-4fe"a 




^j __ CO C 

^j-jCO O ^ CO O^j CO^ 




1 .1 


~ ^ '"^a;^ ^-^S^^ 


03 -- 

w e->5 

W S'>25 '>2w E*> | 


u < 


< <3 c/] 





*0 3 






o Is 







O O 




$ b i 


H U 

S^ g> S^ g> 
. J -1 ^ .2 

'Sag a g 

J gs is 

H 13 H "c3 

w ^.S 



ag 2 

Ei 1! 

So g 


> c co 

K *> 


6 " 

n tN 

j. to r 

o c 

^ I ^ 

O *60 

d -o 

a G 

"2 ^2 


en C -M 

fe 2< 


CU g>W rt 

1 W .1 W < 

s 2 | 6 <" 


1 1 

Ui ip 


CD (S 


M l-t 

M O 

s s 

_r y -> ^ 

q <y 



en en 

^ 3 ^ ! 




? j..O ^ ^ v f^ 



' ' 

5 1 | 1 



U hJ CD IH hJ 



C O 

O 5 


CJ 1 




03 "cu 

Cu ^C^ 

'o i! 

" S 















S bO 

o3 C 

en bo 
o3 C 

oT g 


c 3 3 

C i 

CD rt 

CD ra 





c '5 

^ il 

o d 



































-i "o 












5 3 






~~* "c3 










o> 'C 


J i, 

o3 cy 

o3 cu 




o -S 

^ " 


^ ^ 


'$ ^ 


















> <U 

















f- 1 t3 









U "2 *<y 




IP jy 





*o cr _^ 




O 1 co 





3 '43 en 






















( C 

to bO 


% bo 

03 _G 

o3 C 

% bfl 
03 _G 

w bo 

co bO 

co bo 

03 G 

i ^ 

43 .-3 


O. rr( 

CX "ci 

^ tti 

CX "nj 

a "03 

CU rt 

Pi tti 

CX o3 


D G 

D G 


0) G 

D G 

cu G 

<u G 



bo <u 

bo <U 

bo <u 

tuO Q^ 

G " 

"So fc 

bo 0> 

bo fl^ 

bO 1> 

"bo u 

Of) "* 

CO rt 

c?) rt 

(7) rt 

c/5 rt 

c/3 rt 

c73 rt 

on rt 

c ^ 










H ^ 






fl , 


















- 1 





Poldihutte Kladno, 

L. BraunsSohne,V6lk- 
labriick, Austria. 

Vickers Sons& Maxim, 
Sheffield, England. 

Vickers Sons & Maxim, 
Sheffield, England. 




rv W 







3 K 

Eisenwerk Domnarf- 

Sybry Searles Ltd., 

General Electric Co., 
Pittsfield, Mass., 




































2 1 




2 1 















Use of 

,-M en ^.M co .- w w 



















CO ! 


03 _C 

"a ta 

"So a 

c/5 rt 






a a 


a aj 


43 ' 

_g .2 

4J3 ^_> 
CX aj 

i c js 

O W 


^ c 


N. Y. 


C O 



o S W I W I W 

H ^1 ^ & J3~ ^ ^ 

w oT 



- 5 ri 

as & ^ 



^ gs ^< " 

Q r i 

| r 


- :0 "> - 


43 43 
:Q .5^ 

43 4^ 
:O V 

43 4J 
:O i 




X gjj i 

^ bii u 

x g. i 

x * ^ 

X en i 


-| .s s 

'1 ' " 


'I -S " 

4* b^o *2 

. |-3 s 



53 In tfc 

^2 J S 

04 S J 

1 ^ "s 
s J s 

3 -O ^ 

O^ O3 & 

u o 



co O 

H 1 ^ 

, ^_ tn 



0) | 




1> "S o3 43 

03 "S -r) C 
'u j C 
^ 03 03 42 


^ M 03 42 

03 "3 -r) 
'a ^ a 

CL W ^ 

G .2 *9 T3 
^ C 
42 ^ to oj 








2 c 

. c 



. C 


" a 

"^ a 

"^ a< 

"o ^ 


bfl ^ 

bfl .^ 

bo ^ 




3 "3 

cr -M 


T3 E 

'3 2 

'ft E 
*3 2 
.2" ^ 



'ft E 

.-2 fa' 



















c 3 

03 t C 

rt c 

^ b/3 
03 _G 

U5 bX) 

03 .G 

w bo 
03 G 






ft ctf 

Q. C 

"ft rt 

43 '^2 



^ IH. 



br cu 

be ^ 

be a; 

L^ QJ 

bO V 

C _t^ 

G ii 

c i 

43 . 

c/5 rt 

(75 rt 

(75 rt 

(75 "* 

^H 5 

(75 rt 


5 PQ"* 


|| ll 

I a , 












* aJ 

o G" 

6 G" 

d G 

d G" 

d G" 

*T-> (jj 

U v 

U aj 

U a> 


U <y 

.S T-T 

oy ^ 

o^ kfl 

c^ W 

c# ^ 

oy ^ 

S '53 

N . 

N . 

N . 

N . 




4< 2 ^ 

^ "- U? 

4! -^ ^ 

*5 2 ? 

"^ u? 

S OJ 3 

^H ^ 

<5 <U 3 

^H Q ^ 

^ 0) 3 


<L) S 

43 /V K 


en C 42 

;| I | 


= S | 

.2 E -9 

jr3 c G 

.2 S -g 
= S 1 

o3 O 2 

.2 6 -g 
s S | 

o3 O 2 

b/0 *C r 

i- i E 














1 J-. 


^ o3 

1 t-> 

^ 03 

1 Ui 


1 U 


43 43 

43 43 

43 43 

-C pC 




O G 


"S "c 








O w 
O bo 



P frl 


-M C 

o> w 







r materi 


| I 




*- cti 







1 E 





o ^ 




bo . 





*O 03 




"a g 



g ^ 



H2 o3 



JO ^ 



" C 

3 43 

2" y 



*3 C 

.5* rt 



1 a 















i-o g> 

rO ^" 


O >O 


83 bO 

<U ^ 

S bO 

a ^ 






rt ,C 

o3 C3 

cj C 

Cd CH 






a "ol 

43 '5 

a 03 

43 '13 

8 1 

a 15 











Si o3 






H " 

H "S 

{-H 13 


f-H ra 






2 M 



^ O 



g '3.S 


. o 

"? o 





to 4J rf 

to ^ 

TJ- 10 


y g g 1 


H o'S 


Is *J 














la Marine 
e court,St. 


les -Liege, 



*3 o 

II . 

. ^ 

~ O fe 

w u | 


ks, Waka- 



w " 



Acieries L< 


Ricardo, Ho 


vanger, N 

V- S" 

I 2 


I mperial 
Steel Wor 

i > 













1 U 

.S 3 


I U, 

bo 4> 

bo <!> 
.S 3 

C* | 

C | 

e * 



o fi 

13 03 

3 -a 

: -a 



o <y 

43 43 









u C 


t^ &* 


ON ^ 

O ^ 

n ^ 










1 U 

O O 



g jjj ' 

M <U 


c ^ 


1 1 



5 ^ | 

M ^ 

bfl . 

43 S 

| 1 

." S 

be jy 












^ 1 







o o 

10 H O 






^- ^ 10 















v -- : 



















g '"s-S 

ir ? 2 ** 


h \ 





W M 


s< 11 

s ffi ^ 


t ( 

"O p4 






* 53 


o -^ 




Gelsenkirchner Berg- 
werks Verein Abt, 
Rothe Erde, Aachen, 

Friedr. Krupp A. G., 
Essen, Germany. 
Schneider & Co., Creu- 

sot, France. 
Norw r ay 

Forges de St. Jacques, 
Montlugon, Fr. 








i i 



o *O 

: ^ 

-C *oj 


1 1 


E c/: 

Part Two 



IT has already been pointed out several times that a great 
advantage of the electric furnace over other metallurgical 
furnaces is that it enables the generation of desirable high 
temperatures. Generally speaking this possibility is made use 
of, and work is carried out at higher temperatures than in gas 
fired furnaces. On this account, therefore, it is immediately 
apparent that the materials used for furnace construction, 
especially those parts in contact with the highly heated charge, 
have to meet particularly high requirements. Mistakes in the 
choice of these materials can very quickly bring about trouble 
in the working of the furnace and, under certain conditions, can 
completely stop the operation. 

During the discussion of the different furnaces the materials 
at present used for their construction were mentioned, but it 
still appears advantageous in the following pages to treat these 
materials as a whole, and with special regard to the conditions of 

The first requirement to be demanded is resistance to the 
high temperatures reached in the furnace. This needs, in the 
first place, a high melting point, which is usually measured by 
means of Seger cones. These are named after their inventor 
Seger. They are small three-cornered pyramids made of various 
mixtures of silicates, and are about 2.36" high. The softening 




point, with increasing temperatures, is carefully observed. The 
following table gives the comparison between degrees Centigrade 
and Seger cone numbers. 






































,350 ' 













02 a 



































,500 I 










1, 880 































A material is called refractory if its softening point lies about 
No. 26, and very refractory if the latter is between 30 and 36. 

Without going further into the properties which influence 
the refractoriness of the furnace materials, we will learn the 
requirements they have to meet. 

The first consideration is that in electric furnaces the materials 
are exposed not only to high temperatures, but also to chemical 
influences. Naturally the results of these influences must be 
felt as little as possible. We can give as the second requirement 
great resistance to chemical influences. Unfortunately these 
harmful influences cannot be altogether prevented, and it is 
necessary to reduce them as much as possible. This is partially 
brought about by having the greatest possible density and 
mechanical solidity of the materials. It is immediately evident 
that these properties are of important influence on the durability 
of the furnace masonry and lining, when it is realized that porous 
material offers much more surface for attack by harmful chemical 
influences than one that is dense. Further great density is 


synonymous with good mechanical strength, so that a dense 
material will most successfully withstand the mechanical in- 
fluences due to movements of the electric furnace. 

The appearance of cracks in the furnace walls lead to the 
same bad results as the use of porous material, for they allow 
the harmful chemical influences to penetrate very deep, and 
offer a large surface for attack. This, therefore, brings about 
the requirements of the ability to withstand the influence of 
changes of temperature. These cracks are due to variations in the 
temperature, which bring about expansion and contraction in 
the materials used. 

The requirements for the materials for furnace construction 
are therefore: 

1. Ability to stand high temperatures. 

2. Resistance against chemical influences. 

3. Great density and mechanical strength. 

4. Permanence of form under changes of temperature. 

The materials to be considered are as follows, each of which 
will be taken up separately in the light of the above requirements: 

1. "Schamotte" fire-bricks. 

2. Acid or silica bricks. 

3. Half "Schamotte" or half silica bricks. 

4. Carbon bricks and carbon mixtures for ramming into place. 

5. Basic bricks and basic material for ramming into place. 

6. Mortar. 

"Schamotte" fire-bricks are made from burnt fire-clay 
known as Schamotte or Chamotte, to which unburnt clay is 
added as a binding material. The clay shrinks more or less 
during the burning. The Schamotte must therefore be burnt 
as thoroughly as possible. The more Schamotte in proportion 
to clay is used in the. brick mixture, the less is a strong shrinkage 
to be feared. Moreover, the shrinkage can be partially neutral- 
ized by adding quartz or quartzite which expand during heating 
to the mixture. 

Under all conditions the shrinkage of the "Schamotte" 
brick is to be most carefully kept in mind, because, for example, 
the use of this material in furnace roofs would be disastrous. 


As opposed to the "Dinas" silica bricks, which will shortly be 
described, these fire-bricks have the advantage that they are not 
so sensitive to changes of temperature, and this advantage is 
more marked the less unburnt clay is used in the mixture. 

Their chief importance in electric furnace work is for heat 
insulation purposes, and they are used in this way, for example, 
in the induction furnaces. For furnace roofs they are only 
applicable if the temperature attained is not very high, and 
therefore at the most can be used only for induction furnaces. 

Acid or silica bricks are greatly used in electric furnaces for 
roofs. They are very rich in silica, and are made from quartzite 
with 95 to 99% silica, and ought to have at least 95% when 
finished. They are known to the trade as English "Dinas" or 
Lime Dinas bricks. For the Lime Dinas bricks an addition of 
i to 2% of limes is used as a binding agent, usually in the form of 
cream of lime. If clay is used instead of lime the quality is not 
quite so good, the bricks containing 80 to 90% silica, and being 
known as Clay Dinas or German Dinas bricks. 

As has been mentioned, these bricks expand considerably 
with increasing temperature. This is less noticeable in bricks 
made from certain quartzites, but is unavoidable even with the 
best materials. Their greatest use is for the roofs of furnaces 
with high roof temperatures, such as all the arc furnaces have, 
and because of their expansion with heat a very flat roof can be 
maintained. Unfortunately these bricks are very sensitive to 
changes of temperature, and offer only small resistance to the 
action of slags. They are therefore practically restricted to roof 

The half Schamotte or half silica bricks are between the 
two kinds of brick just described. They consist of a mixture of 
quartz and burnt clay, and have properties corresponding to an 
excess of one or the other constituent. 

Carbon bricks and carbon mixtures for ramming into place 
are not used in steel-refining furnaces; because of the great 
affinity between iron and carbon they are quickly destroyed, 
and undesirable carbon enters the metal. On the other hand 
carbon mixtures have been often used in electric shaft furnace 


experiments in their double capacity as refractory materials and 
conductors of the current. Apart from this, carbon bricks or 
mixtures are not used for electric furnace construction. 

Basic bricks and materials, however, are so important that 
they are used almost exclusively for the hearths and walls of 
electric steel-making furnaces. They include chrome iron ore, 
or chromite, dolomite, and magnesite. 

Of these materials chromite is not used in Germany. It 
has the disadvantage that if it comes in direct contact with the 
metal it is rapidly destroyed, and influences the bath in an 
unwished-for and harmful way. 

Dolomite is finding increasingly large application. It is a 
limestone with a large percentage of magnesia, CaC0 3 MgC0 3 , 
and is found in large amounts in Thuringen and Lothringen, 
and in various parts of the United States. Its greatest use in 
the iron and steel industry is as a lining for the basic Bessemer 
converter for which it is prepared in so-called dolomite plants. 
The method of preparation may be briefly described: 

The raw dolomite is either broken by hand, or crushed, to 
pieces about the size of one's fist. These pieces are then burnt 
in a shaft furnace to a clinker. The amount of coke necessary 
is from 20% to 30% of that of the raw stone. After burning, the 
clinker is ground in a suitable mill, the largest pieces being not 
more than 0.4" diameter. The ground dolomite is then mixed 
with about 7 to 10% of hot dry tar, on a moderately heated 
floor. This mixing is carried out either by hand or in a chili 
mill, or suitable mixing machine, which is so constructed that 
the material can be heated. Careful attention must be paid to 
the preparation of the tar if the basic dolomite-tar mixture is 
to have the best properties. The crude tar is distilled in special 
apparatus at 240-280 C., and is freed in this way from the 
ammonia water and light oils. It must be mentioned that the 
burnt dolomite, because of its large lime contents, readily absorbs 
moisture from the air and falls to powder. It should, therefore, 
be used as soon after its preparation as possible, and is applica- 
ble to the making of bricks, or for ramming into place. 

Magnesite, as well as dolomite, is finding an increasingly 


large use for those parts of the furnace in direct contact with the 
charge, and indeed for all those parts exposed to specially high 
temperatures. It has an advantage over dolomite in that it is 
more neutral in character. When mined it has a melting point 
about equal to Seger cone, No. 42. As it shrinks a great deal 
when heated, it must, before using, be burnt so thoroughly that 
it is sintered. For this a temperature of about 1700 C. is 
necessary, and the~Sp. Gr. rises from 3.19 to 3.65. If the magne- 
site is to be used direct, it is mixed with tar, like the dolomite, and 
rammed into place. If, however, it is to be made into bricks, no 
tar is used. The finely powdered magnesite is forced into shape 
by the use of very high hydraulic pressure, and then again burnt. 

Apart from its greater neutrality magnesite has the further 
advantage over dolomite that it is not so sensitive to the action 
of moisture, and when burnt to sinter can be kept in storage 
without fear of spoiling. On account of its great density, how- 
ever, it easily cracks when subjected to changes of temperature, 
but does not shrink very much with increasing temperature, and 
because of this is being used more often for furnace roofs. 

When building electric furnaces a suitable mortar must be 
used. Here again permanence of form under changes of tempera- 
ture is of the first importance, and the mortar must be carefully 
considered in this respect. Also the courses should be laid as 
close as possible, so that even with expansion of the mortar the 
joints may remain tight, and expose very small surfaces of at- 
tack to harmful influences. As a general rule it may be said 
that the mortar should be of the same nature as the rest of the 
construction, using basic mortar with basic materials, and acid 
with acid. Thus with dolomite bricks either tar alone is used, 
or tar mixed with dolomite as a binding material, while with mag- 
nesite bricks powdered magnesite is used mixed either with tar 
or a little hydrochloric acid. With other than basic brickwork 
a certain amount of quartz or white silica sand is usually mixed 
with the mortar, an excess of clay being avoided, and care again 
being taken to have tight joints. 

Reconsidering the materials of construction we see that none 
of them possesses completely the qualities that have been men- 


tioned, so that they must be carefully chosen and used with 
proper regard to their properties. This leads us to a point that 
must not be left without attention. It is that, under the influence 
of other materials at high temperatures, certain refractories 
quickly break down. The following rule must be strictly 
observed : In the presence of basic influences use basic materials, 
and with acid influences use acid materials. Therefore, in a 
furnace in which the basic process is to be carried out, basic 
materials must be used completely or at least in those parts 
where the temperature is high enough for one material to act 
on the other. For instance not alone the hearth and walls but 
the cupola-like roof of the Stassano furnace is built of basic 
material, the roof being of magnesite bricks, and the hearth of 
either dolomite or magnesite bricks, or these materials mixed 
with tar and rammed into place. On the other hand, in the 
Heroult and Girod furnaces the hearth is basic but the roof 
is of silica bricks. This is only possible when no danger of 
fusion is to be feared at the junction of the two materials. It 
should be remembered that quartz and clay, or mixtures of the 
two, so greatly lower the melting points of dolomite or magnesite 
as to bring them from Seger cone 17 to n, which would naturally 
lead to the rapid destruction of any electric furnace if it took 

The extremely harmful occurrences mentioned here must be 
prevented, either by the use of basic material exclusively, or 
else the contact between the acid roof and basic hearth must be 
removed from the influence of harmful temperatures. 

It may be mentioned here that the lowering of the fusion 
point of a refractory material is often brought about purposely, 
however only to a certain degree. For instance, in making a 
hearth of magnesite, slag, clay, or similar material is often added 
as a flux, to bring about a more ready sintering of the mass. In 
this way the hearth is made denser, and offers greater resistance 
to the influence of the metal and slags. Naturally only the 
smallest amount of flux necessary must be used in order not to 
reduce the melting point too much. 

In the design of electric furnaces the properties of the ma- 


terials to be employed should be carefully kept in mind. In 
the first place the hearth ought to be of such a shape that, after 
a charge or a short run, repairing can be carried out. This is 
necessary because of the unavoidable action of the slag on the 
lining. This is completely possible at present only in that 
furnace which is patterned after the open hearth, namely, the 
Heroult. All other furnaces allow a certain amount of relining, 
but because of the vertical or almost vertical walls it is impossible 
to keep the hearth longer than about three weeks. After this 
time a new one is necessary, equally with the Stassano, Girod, 
or Rochling-Rodenhauser. Notwithstanding this, the previous 
discussions on costs of operation have shown that the lining costs 
of the Heroult and Rochling-Rodenhauser are about the same. 
This is because of the more rapid destruction of the roof in the 
arc, as opposed to the induction furnace. 

It is because of these considerations that the Stassano and 
Girod furnaces do not have such good lining costs as the Heroult 
and Rochling-Rodenhauser. In case bricks are not used the 
new hearth is made up of hot magnesite or dolomite mixture 
rammed into place. If this work has been carried out by hand 
it can be more profitably done by means of compressed-air 
hammers which are sold by all the firms making compressed-air 
tools. Air at a pressure of about six atmospheres is used. These 
tools bring about a great saving in labor, and have the ad- 
vantage that they also give a much denser and more solid hearth 
than hand ramming. Naturally if such a lining is heated, which 
corresponds somewhat to the burning of refractory bricks, an ex- 
pansion takes place. This will certainly give rise to cracks unless 
a certain freedom of movement is allowed for. This is well pro- 
vided for by leaving a space between the hearth and the insulating 
cover, which may be filled with loose granulated material. 

In all arc furnaces the roof should be made removable, for 
repairs are not possible during the operation. In this way a 
new roof can easily and quickly be put into place. As the 
bricks used for the roofs are sensitive to changes of temperature, 
it is evident that there will be more danger, the more water- 
cooled openings there are through which to pass electrodes. In 


regard to further details concerning the construction of the 
various types of electric furnaces, reference must be made to 
the chapters of the earlier part of the book, in which the furnaces 
are separately considered. 


The question of the operating costs of electric-steel and 
pig-iron processes is undoubtedly the most important one after 
that of the quality of the product. These two questions there- 
fore will determine which types of furnace will advance in the 
future, and which will recede. With regard to the quality it is 
generally acknowledged that all the accepted types of electric 
furnaces, those considered in detail in the chapters of the first 
part of the book, will produce steel that will answer all require- 
ments. However, the possibility is often mentioned of an un- 
favorable influence on the quality of the steel of the unnecessarily 
high temperatures of the arc furnaces. This possibility is 
brought up again and again, and was recently spoken of by 
Henry M. Howe, Professor of Metallurgy in Columbia University 
(E. & M. J. } Aug., 1909). 

The certainty of different methods of heating causing a 
difference in quality can only be proved if two steels are made 
of the same composition, one in an arc and the other in an induc- 
tion furnace, the same charge to be used in each case, and 
equal expert supervision given; then the two steels to be 
thoroughly tested to see whether they give different results. 
Unfortunately, such a test has not yet been made. 

Until this is done, we must admit that steel made in the 
different furnaces, with expert supervision of the metallurgical 
processes, has the same excellent properties. There remains, 
therefore, only the question of the operating costs to determine 
which type of furnace to adopt and the economy of electric steel 
production. Calculations of the operating costs of the different 
types of furnaces have been published many times, so that it 
would appear very simple to compare them one with the other. 
This would immediately show which one would allow the cheapest 


production of steel, and the difficult question of the choice of the 
most suitable and economical type of furnace would be solved 
at one blow. Unfortunately this method is altogether incapable 
of giving a view corresponding to the real conditions. The figures 
that one usually finds published are often misleading. However, 
some of them should be considered later, for with proper care 
they will give much interesting information. 

First we will see which factors are of importance in influenc- 
ing the operating costs. 

They are briefly: 

1. The materials charged. 

2. The loss during the operation. 

3. The consumption and price of the current. 

4. The fluxes. 

5. The labor costs. 

6. The costs of linings and repairs. 

7. Amortization. 

8. Costs of electrodes. 

9. Auxiliary appliances. 
10. Tools. 

In regard first of all to the material charged, it can be of 
metal below the average in quality, and therefore cheaper, for 
all electric furnaces which are suitable for refining. This is be- 
cause these furnaces allow a very complete removal of all harmful 
impurities. The induction furnaces with ring-shaped hearths, 
such as the Kjellin furnace, do not allow thorough refining to 
be carried out, as has been pointed out in the previous chapters. 
For these furnaces, therefore, especially pure and correspondingly 
dearer raw materials must be chosen, similar to those now used 
in the crucible process. This includes the high-priced Swedish 
irons, as well as the purest refined metal, usually made from 
Styrian charcoal pig iron; also fluid metal already refined in 
the open-hearth furnace, which can be " killed" in the electric 
furnace, in exactly the same way as in the crucible. Such a 
furnace is therefore at a decided disadvantage with regard to 
the metal charged compared with other furnaces, notwithstanding 
that a considerably lower power consumption is naturally re- 


quired for the further working up of pure raw material compared 
with material that must be first refined in an electric furnace of 
another construction. 

For these electric refining furnaces, the cost of the charge 
can be taken as equally high when comparing the operating 
costs. When comparing the costs of the open-hearth and electric 
furnace it must not be left out of consideration that the latter 
has the advantage that it allows the use of more impure and 
therefore cheaper raw materials, at the same time permitting the 
production of the highest quality steels. 

It appears unnecessary to give here any figures on the price 
of scrap, for as already mentioned this price is strongly depend- 
ent on local conditions. This also applies to fluid charges from 
either the blast furnace, cupola, open-hearth furnace, or converter. 


This means the material lost during the treatment of the 
molten metal in the electric furnace. Compared with other 
furnaces it is very small. When using the purely induction 
furnace with ring-shaped hearth, in which the purest material 
must be used, the loss is not considered at all; it can be put down 
as zero. On the other hand, when carrying out refining in the 
electric furnace, a certain loss is unavoidable. Altogether apart 
from the slagging of the impurities, small amounts of the liquid 
metal are torn away when the slag is removed. This loss is 
therefore the greater the more impurities are present in the 
charge which necessitates a more frequent making and removal 
of slags. One will scarcely make a mistake in taking this loss, 
depending on the charge, as about the same in all electric refining 
furnaces. It should be figured on the average as 4 to 6 per 
cent, with a solid charge, and 2 to 3 per cent, with a liquid 

In this connection it must be remembered that very light 

*Borchers gives the loss as 10 to II per cent, in a Girod furnace with a cold 


thin scrap can be used in the electric furnace such as waste wire, 
turnings, etc., without the loss being higher than the figures 
given above. This is because the strongly oxidizing action of 
the hot gases of the open hearth is not present in the electric 
furnace. For comparison it may be mentioned that the loss in 
the open-hearth scrap process ordinarily amounts to 4 to 8 per 
cent., and is considerably higher if much of the light scrap, 
mentioned above, is used. 

If a pig-iron process is carried out in the electric furnace, 
which can offer an economic advantage under very favorable 
prices for current, then more iron is reduced from the ore than is 
the case in the open hearth, so that the yield is easily greater in 
amount than that charged. In this respect the electric furnace 
works more cheaply than the open hearth, and further allows the 
production of a higher quality of steel than can be produced 
in the open hearth. This is shown very plainly in that steel is 
often taken from the open hearth to the electric furnace to be 
refined, or to be alloyed, etc. 

In view of these advantages, why is not the electric furnace 
used more often for melting, in place of the open hearth? The 
answer is found immediately if we consider the cost of heating, 
on the one hand in the open-hearth or crucible furnace, on the 
other hand in the electric furnace. The fuel consumption in the 
open hearth, working the scrap process and using bituminous 
coal, amounts to 22% to 32% of the output. If we take the 
highest value, and remember that the electric furnace in the 
sizes used up to now can only be compared with small open- 
hearth furnaces, we find that 705.5 Ibs. of coal would be used 
per metric ton of open-hearth steel. In the electric furnace 
750 to 800 kw. hours would be necessary to melt the scrap used 
in the open hearth, and to refine it to the same grade as ordinary 
open-hearth steel, provided that the furnace was of 5 to 8 tons 
capacity. If the coal cost $3.57 per ton, then the heating cost 
alone of the open hearth would be $1.12 per metric ton. In 
order that the heating cost in the electric furnace should not 
exceed that of the open hearth, the kw. hour, with the above 
assumptions, should cost o.i5c. 


If the kw. hour prices are calculated, which are allowable 
with different prices of coal and coal consumption, so that the 
heating costs in the electric furnace do not exceed those in the 
open hearth, then the following table is obtained: 

Used per 
Ton of 

Allowable Price per Kw. Hr. in Cents with Coal at the Following 
Price per Metric Ton (2,204 Lbs. ): 

If there 
is Used 
Ton of 











. Ill 


. 140 


. i66c. 




. 121 






. IOO 




. 164 







. 140 


.I 7 8 





114 -133 







. 121 








.078 .090 

. 104 







. IOO 

















. IOO 






























. 1470. 





. 121 







. 102 











. I4O 





. IOO 















This table clearly shows how very much cheaper the heating 
costs are in the open-hearth than in the electric furnace. With 
the most unfavorable coal consumption (32%), and very high cost 
of coal ($5.71 per metric ton), the kw. hour ought not to cost 
more than o.238c. If we calculate the kw. year as containing 
300 working days, then the kw. year ought not to cost more 
than $17.14, or the e.h.p. year must not cost more than $12.62. 
It is evident that these prices for power can only be reached 
with the most favorable conditions, for example through the 
use of water-power. It is therefore also clear that the electric 
furnace can only be used for the melting of scrap, and the pro- 


duction of steel, similar in quality to ordinary open hearth, in 
such places where the cheapest natural power is ready for use, or 
else where small amounts of steel are to be made for which the 
open-hearth process is unsuitable. In all other cases it is almost 
always preferable to leave only the final work to the electric 
furnace. In this way at only a small increase in cost an improved 
quality is reached, compared with open hearth which is greatly 
in favor of the electric furnace. 

It has been mentioned already that the electric furnace can 
replace the crucible. If we therefore now consider the heating 
costs of the electric furnace on the one hand (using this method 
of working), and the crucible on the other, we obtain the follow- 
ing: The fuel consumption with crucible melting in coke furnaces 
:not using the waste gases amounts to about 150 to 200 per cent. 
<of coke, and with the use of regenerative gas furnaces, 175 to 200 
per cent. coal. 

If we take it, first, that the same pure charge is to be used 
in the electric furnace as in the crucible, then the melting capacity 
.alone of the electric furnace comes into consideration. For this 
the power consumption is 600 to 750 kw. hrs. per metric ton of 
finished steel depending on the size of the furnace used. For 
instance, the firm of Krupp has brought the power consumption 
down to 617 kw. hrs. per metric ton in their 8-ton, ring-shaped 
induction furnaces, of the Kjellin and Frick types. Also, calcu- 
lating on the melting alone, the power consumption in an 8-ton 
Rochling-Rodenhauser furnace is only 580 kw. hrs. per ton. 

Taking the figures given above as a basis, the table on the 
following page clearly shows how high the cost per kw. hr. 
may be, in order that electric heating may not be dearer than 
that in the crucible furnaces with the given unit prices for coal 
and coke. 

The table shows how the unit price for power can increase 
very considerably before the heating cost in the electric furnace 
will exceed that of the ordinary crucible furnaces. We had 
assumed previously that only pure raw materials were used. 
In this way, however, only a part of the advantage of the electric 
furnace is utilized since the remarkable properties that it shows 


as a refining furnace remain unemployed. We should obtain r 
therefore, a still more favorable idea of the electric furnace if 
we used less pure and, therefore, cheaper raw material in the 
charge. Although from 150 to 250 kw. hrs. more would be 
required, depending upon the degree of purification, this increase 
would scarcely outweigh the savings brought about by the use 
of a cheaper charge. 


Used per 
Ton of 
Per cent. 

Allowable Price per kw. hr. in Cents with the Following Coal 
or Coke Prices (per Metric Ton) : 

If there 
is Used 

Ton of 
kw. hrs. 























I. Ill 





> 600 




i. in 













1. 100 













t 6 50 

















1. 02 1 













t 700 


























1. 000 

1. 112 

1. 22 I 


t 750 











It must be further considered that much more labor is 
required to operate the crucible furnaces than an electric furnace, 
which can replace many crucibles because of its capacity. This 
latter property brings about a further advantage, namely, a 
complete uniformity of the whole cast, while the material from 
different crucibles shows certain variations. It should also be 
mentioned that the cost of crucibles is higher than that of the 
upkeep of an electric furnace. Finally when one considers 
that the steel from the electric furnace is of fully equal value 
to that from the crucible, then the displacing of the crucible 



by the electric furnace appears inevitable. This is shown by 
the growth that the electric-furnace industry has had even up 
to now. The following figures in metric tons are taken from the 
steel production of Austria-Hungary: 


Crucible Steel 

Electric Steel 


2"? 21^ 













We have previously shown that, in regard to heating costs, 
the electric furnace is more economical in almost all cases than 
the crucible furnace, but that on the other hand it usually is 
less economical that the open hearth. This naturally brings 
it about that as much as possible of the melting and refining 
should be done in the more cheaply operated open hearth; or, 
in the case of the refining of basic Bessemer metal, in the con- 
verter. This leaves only refining and desulphurization for the 
electric furnace, for both of which purposes it is particularly 
suitable, because of the easy regulation of the temperature, and 
the removal of the harmful influences which are unavoidable 
with any other method of heating. 

It is, therefore, to be expected that the electric furnace will 
not only displace crucible plants, but will be introduced more 
and more in connection with open-hearth and Bessemer plants. 

The power consumption necessary for the work of refining 
naturally depends greatly on the final product desired, but it is 
also dependent upon the degree of purity the material has, when 
charged into the electric furnace. Furthermore, the size of the 
furnace, as well as the efficiency of the particular type of furnace 
chosen, has an influence which must not be neglected. In re- 
gard to these latter influences, the discussion in the first part 
of the book must be consulted. 

The only points remaining to be considered are those of the 
material charged, and the final product required. 


It is known that in melting in any furnace a higher efficiency 
is obtained the quicker the melting proceeds, that is, the greater 
the amount of energy supplied, the higher the efficiency. With 
an important lessening of the time necessary for melting, there is 
a corresponding lowering in the amount of heat lost by radiation, 
etc. It is also well known that, after the melting stage is once 
over, the following refining period cannot be lowered at will by 
increasing the amount of energy introduced, but that this 
refining work requires a certain time. As already mentioned 
several times, the slag must be changed more frequently depend- 
ing on the impurity of the charge and the required purity of the 
final steel. The curve given in Fig. 60 (see Part I, page 134), 
which shows the power consumption depending upon the size 
of the furnace with different slag changes, gives a fitting idea 
of the influence of the impurities in the charge on the power 
consumption. The figures given should therefore be considered 
as approximate. To give more exact values is apparently only 
possible with a thoroughly fixed type of furnace, of a fixed size, 
and with an exactly established charge and final material. 

For example, basic Bessemer metal with about 0.08% P 
and 0.08% S, requires an average of 250 kw. hrs. per metric ton 
for refining, in an 8-ton Rochling-Rodenhauser furnace, when 
the final material required is of crucible steel quality with a 
definite carbon content. With the production of the highest 
value alloy steels the power consumption under almost the 
same conditions increases to 280 and even 300 kw. hrs. per 
metric ton. On the other hand, when making structural steels 
it falls to 200 kw. hrs. or less. The power consumption is there- 
fore the smallest when only a limited alloying or degasification 
must be carried out, and not a thorough refining of the metal. 
It then falls even to 100 kw. hrs. and less per ton. 

It should be remembered that very impure metal was taken 
for the charge, basic Bessemer, and if metal was taken from the 
open hearth for example, with 0.03% P and 0.05% S, then 
under the same conditions there would be a certain lowering 
of at least 50 kw. hrs. per metric ton when making high quality 


The considerations given above serve to show that the 
power consumption figures given in technical papers should 
be carefully investigated to see what conditions they refer to, 
for such figures only lead to grave mistakes in many cases. 

As the electric pig-iron furnace is beginning to be of impor- 
tance, as shown by the action of the Jernkontoret in Sweden, who 
have built a furnace for a daily output of 20 tons with an energy 
consumption of 2500 to 3000 h.p., a comparison is given below 
between the ordinary blast and electric shaft furnace. The 
following table, due to Catani, is taken from Neumann's paper 
in Stahl und Eisen, 1909, p. 276, jf. It shows what unit 
prices may be paid for electrical power so that the heating cost 
in the electric furnace does not exceed that of the ordinary 
furnace, with the given price of coke and output per h.p. day: 

Weight of Pig Iron in 24 Hours 

Allowable Price per H.P. Year with the Following 

per H.P. 

Coke Prices r 











I 7 .6 



12. 19 



7 .6l 

II .42 







By calculation we obtain the following table, taking the 
h.p. year as equalling 0.736 kw. year, and the year as containing 
365 days: 

Weight of Pig Iron in 24 Hours 
per Kw. 

Allowable Price per Kw. hr. in Cents with the Following 
Coke Prices: 





















I6. 3 





In order to be able to form an opinion from the figures given 
in the table, it is naturally necessary to know what efficiency is- 


possible today with the electric pig-iron furnace, per kw. day. 
This naturally depends in the first place on the quality of the 
ore used. In the following chapters the metallurgical part of 
this work will be gone into, but the table already shows that with 
coke at a relatively high price, the price for electricity must 
be very low for the electric-shaft or pig-iron furnace to compete 
with the ordinary one. In Germany, therefore, the electric- 
shaft furnace apparently has no future. This is clearly shown 
in the following table by Neumann (Stahl und Risen, 1904, p. 
143). Here the carbon necessary for the reduction of the various 
ores used in Germany, and that replaceable by electric power is 
calculated and given in money value. The price of coke is taken 
as $3.57 per metric ton, and that of power as o.238c. per kw. hr., 
or $19.04 per h.p. year. 


The Carbon Replaceable by Elec- 
trical Energy 


Iron Ore 

Pig Iron 





Cost of 


sary for 

by Kw. hr. 






Bilbas brown 


722.9 Ib. 

1197. i Ib. 

1400.5 Ib. 





iron ore 


327.9 kg. 

543 kg. 

635.3 kg. 



910.5 Ib. 

1247.8 Ib. 

1460.0 Ib. 







413-0 kg. 

566. kg. 

662.3 kg. 

red iron ore 



509.3 Ib. 

1261.0 Ib. 

1475.7 Ib. 





burg Loth- 


231 .0 kg. 

572. kg. 

669.2 kg. 

ringen Min- 




1067.0 Ib. 

806.9 Ib. 

944 . i Ib. 





M a gne t- 


484.0 kg. 

336. kg. 

428.2 kg. 


The next question is: What unit prices for electrical power 
are obtainable today? This has been treated already in Chapter 
XV of the first part of the book, and it is therefore sufficient to 
give here merely the figures on which rough calculations can be 
based with the use of water-power o.ii9C. and more per kw. hr. 
and more. 

With the use of blast-furnace gas-engines 0.3570. to 0.7140. 

Steam turbines of great efficiency 0.7140. and more 

Steam-engines 0.9520. " " 

Overland and large city central stations 0.9520. ' 

These figures show the values reached under the most favor- 
able conditions. Apart from these, the prices naturally depend 


very largely on local conditions, so that for more exact calcula- 
tions these conditions must be considered. Further, the figures 
refer to the delivery of power at the generators, so that for 
exact calculations, the transmission losses, and losses in stationary 
or rotating transformers must also be considered. In the latter 
case, for example, these can easily amount to 20%, so that the 
cost of power at the furnace is 20% higher than at the central 

We have now sufficiently considered the influence of current 
consumption and cost on the operating costs, and can pass on 
to the other points. 

The fluxes necessary for the operation of electric fur- 
naces depend in the first place on the amount of the impu- 
rities in the charge, and further on the desired composition of 
the final material. Also, on the method of carrying out the 
refining process, or on the furnace construction or method of 
heating, which under certain conditions may bring about a 
special method of working. As has been pointed out in previous 
chapters, lime and roll scale or ore are necessary during the 
oxidation stage. During the deoxidation stage, more lime, 
together with some sand or fluor-spar, are used to make it liquid, 
and some powdered carbon or ferro-silicon as special deoxidation 
medium. Carbon is used only in the Heroult furnace, all other 
arc furnaces and also the Rochling-Rodenhauser using ferro- 
silicon, so that in these latter furnaces a somewhat higher ferro- 
silicon consumption has to be figured upon than in the Heroult 
furnace. Further, all furnaces working with carbon electrodes 
use a slightly greater amount of oxidizing agents during the 
oxidation period compared with induction furnaces, due to the 
reducing action of the carbon vapor. This must be reckoned 
with, altogether apart from an increased power consumption.* 

The wages or labor costs which are required for the operating 
of electric furnaces, calculated per ton of steel, are the smaller 

* That an increased power consumption is required for arc furnaces com- 
pared with induction furnaces, due not only to the reducing atmosphere in 
arc furnaces but also because of the greater electrical loss, was proved in the 
first part of the book. 


the greater the capacity of the furnace and the larger the amount 
of steel produced. In almost all cases the labor necessary to 
operate a small furnace will be completely satisfactory to operate 
a larger one. If we consider that the size of the various types 
of furnaces is the same, then the labor necessary for the purely 
metallurgical work can be taken as equal in amount. It should 
be determined whether solid or liquid charges are to be worked, 
and in the latter case the kind and amount of scrap to be charged, 
as well as the kind of auxiliary machinery to be used. If we also 
suppose that the number of men necessary to handle molds and 
work on ladles is the same under all conditions, for the different 
furnaces, (which appears to be absolutely correct,) then the 
same amount of labor would be used with all the furnaces for the 
purely metallurgical work. 

We have already noticed, however, in the first part of the 
book that with the Stassano furnace one man is necessary to 
watch continuously the electrical recording instruments, and 
to regulate the electrodes according to their readings. Such a 
man is necessary with all arc furnaces unless they are provided 
with automatic regulating arrangements, and even if these are 
present a continuous supervision of the electrical conditions is 
necessary while the scrap is being melted, for example in the 
Heroult furnace, as has been already pointed out in Chapter VIII. 
This extra man is unnecessary with induction furnaces, and 
with proper design of the furnace all the switches and regulation 
devices can be looked after by the first melter without any great 
or important waste of time. 

When working with fluid charges in arc furnaces equipped 
with automatic regulation no important switching work is neces- 
sary, and the special expense can be saved. These conditions 
are not without bearing on the amount paid for labor per ton of 

The lining and repair costs form a very important part of 
all operating costs. They include labor and the expense of 
material. The material costs, in the first place, depend largely 
on local conditions so that correct unit prices cannot be given. 
Apart from this the wear and tear on the furnace roof and walls 


depend very largely on the method of heating. For this reason 
we find, for example, that the roof is strongly attacked in all 
arc furnaces, as it is exposed to the heat radiated from the arc, 
while an attack on the roof of induction furnaces can scarcely be 
noticed. The reason is that in the latter case the heat is pro- 
duced in the metal bath itself so that the roof is protected by 
the covering of slag, altogether apart from the fact that at no 
place is a temperature of 3500 C. produced, as is sometimes the 
case near the carbon electrodes. 

In all electric furnaces there is also a certain wearing away 
of the dolomite or magnesite hearth by the slag. As long as 
possible this is taken care of by repairs made between the charges. 
This is done the more easily if all parts of the hearth can be 
reached from the doors, and if the material used sticks to the 
places to be repaired. The Heroult furnace has the best shape, 
while the more cylindrical Girod and Stassano furnaces, as well 
as the Rochling-Rodenhauser only allow such repairs to a 
certain extent, so that after a run of a certain number of charges 
the furnaces must be stopped for repairing the walls, and in the 
case of the Girod and Gronwall, the bottom also. This brings 
about a certain loss of time and 'expense for labor, both of which 
are the greater depending on the difficulty of making the walls 
and roofs. The Stassano shows the most unfavorable conditions 
in this respect, while the Girod and Rochling-Rodenhauser 
can be prepared for operation in about the same time. With 
the latter a new lining is necessary after each 100 to 120 heats. 

In regard to furnace repair costs it is evident that with 
arc furnaces the price of material for the roof as well as the 
hearth is of determining influence, while for, induction furnaces 
the latter alone is of special importance. In general it may be 
said that the repair and maintenance costs of the furnaces mostly 
used, namely the Heroult, Girod, and Rochling-Rodenhauser, 
do not exceed those of the open hearth, as soon as heats averaging 
3 tons and upwards are worked. In open-hearth furnaces this 
can be taken as 36 to 6oc. per ton. 

The depreciation is naturally higher, the more expensive the 
whole plant having the same capacity. It is therefore important 


to use the plant as completely as possible, and the induction 
furnace undoubtedly allows this the most easily, as it works 
without rapid current variations. As this furnace moreover 
has undoubtedly the best working efficiency, and can be kept 
under current continuously, even during charging, without 
the machinery being in danger, there is a saving in time, 
and therefore an increase in production for a given size of 

With an equal amount of total plant cost the depreciation per 
ton of steel with the induction furnace must be smaller than with 
other electric furnaces. In regard to the extent of the cost of 
plant itself, the first part of the book may be referred to. 

Electrode Costs. This comes into question only with arc 
furnaces. The conditions affecting the consumption of electrodes 
were treated in Chapter VI of the first part of the book. It 
was also proved in Chapter IX that the Girod and Heroult 
furnaces should be considered as working with the same electrode 
conditions, provided that both furnaces are technically of the 
same excellence. We can, therefore, without further thought 
put down the electrode consumption in these two furnaces as 
equally high. On the other hand the Stassano furnace, working 
under altogether different conditions, will give another electrode 
consumption. The electrode material will also naturally affect 
the cost per ton of steel. Carbon electrodes vary in price from 
$5.95 to $9.52 per metric ton; graphite electrodes, $15.47 to 
$53.20. Carbon electrodes can sometimes be produced con- 
siderably cheaper in one's own plant, but this presupposes 
very large electrode consumption and a very large electric 
furnace plant, otherwise the cost of one's own electrodes will be 
higher than that of those from a special plant. 

It is perhaps not without value to consider that the mild 
steel pole pieces, such as are used in the Rochling-Rodenhauser 
furnaces, are not attacked. As is well known they are protected 
from the high temperatures of the bath by a conductor of the 
second-class, which is composed of the lining itself. Through 
this arrangement every electrode cost disappears. 

Certain operating costs proceed from the auxiliary machinery 


necessary with all furnaces. For instance, with the Rochling- 
Rodenhauser furnace there is the air cooling of the transformers, 
and with all arc furnaces a certain water consumption for cooling 
the electrodes, or for the governing of the electrodes as in the 
Stassano furnace. To this also belong the costs of the power 
necessary for the tilting or turning of the furnaces, and finally 
also that necessary for automatic regulation, etc. These costs 
altogether are, however, only very small. With all electric 
furnaces they only amount to a very few cents per ton. Finally 
a certain consumption of working tools, rabbles, rods, etc., should 
not remain unmentioned, which should cause about the same 
costs for all furnaces. Also when calculating the costs exactly, 
the power for lighting, operating the travelling cranes, etc. r 
should be considered, which can be taken as equally high for the 
different furnaces. Finally, there is a license cost which comes 
into question, concerning the amount of which only the companies 
owning the patents can give information. 

As a conclusion some operating costs may be given for 
different furnaces. It should be again pointed out that such 
figures and comparisons are to be used with the greatest care 
because they are based altogether on local conditions, and also 
on the kind of metal charged and obtained. In regard to the 
operating costs of the electric shaft furnace it has been pointed 
out already that it can only compete with the ordinary blast 
furnace under the most favorable conditions. These conditions 
exist, for instance, in some parts of the United States, Canada, 
Norway, Sweden, and Switzerland, and the following comparison 
of costs is for Sweden. 

It has been made by Prof, von Odelstierna of Stockholm, and 
is taken from the Electro Chemical and Metallurgical Industry, 
1909, p. 420. 

In the charcoal blast furnace: 

0.950 metric tons charcoal at $8.00 per ton $7 . 60 

Labor i . oo 

Repairs and general expenses 1 . 50 

Per metric ton $10.10 


In the electric shaft furnace: 

o. 270 metric tons charcoal $2.16 

0.3 electric h.p. years at $12.142 3.60 

Labor i . oo 

10 Ib. electrodes at 3c. Ib 30 

Repairs and general expenses i . 50 

Per metric ton $8 . 56 

According to these figures the use of the electric furnace 
gives a gain of about $1.54. It is based on the assumption that 
the Gronwall, Lindblad & Stalhane furnace, which has shown 
up the best so far, is taken as the electric shaft furnace. The 
ore taken for the comparison ought to contain 60% metallic 
iron, and the charcoal 83% carbon. 

It is further assumed that both furnaces have the same 
output from 8000 to 10,000 metric tons per year. The prices 
for ore and raw limestone* are not taken into consideration, as 
they depend so largely on local conditions. 

If we compare the results found here with the previously 
given table of the cost of current for the electric blast furnace, 
we find complete agreement. For instance, from the table on 
page 230, we see that if the cost of heating in the two types of 
furnace is to be equally great the h.p. year should cost $12.18. 
This is with a production of (8 kg.), 17.637 Ib. pig iron per h.p. 
day, and a price of coke of $7.61 per metric ton. The figures 
of Prof, von Odelstierna are based on power at $12.14 P er h.p. 
year, charcoal at $8.09 per metric ton, and an output of i metric 
ton per 0.3 h.p. year. This corresponds to about (9 kg.), 19.841 
Ib. per h.p. day. If it is assumed that the coke and charcoal 
contain the same carbon then the estimate of von Odelstierna is 
calculated with a higher output and with a greater price for 
carbon, both of which points are favorable to the operating costs 
of the electric-blast furnace. 

It should, however, be again pointed out that such favorable 

* In the Metallurgical and Chemical Engineering, Feb., 1912, p. 71, 
LEFFLER says that in practise it has been found more economical to use 
unburned limestone, and that among other things burned limestone causes 
the burden to hang. 


conditions for the electric-shaft furnace are not often present, 
so that it is restricted to countries poor in fuel and rich in ore 
and electricity. 

In this respect it is, however, encouraging to note that, after 
the five months' test made at Trollhattan, ending April, 1911, 
(according to The Iron and Coal Trades Review, of Nov. 10, 1911), 
the pig iron produced per h.p. year equalled 3.79 metric tons or 
22.92 Ib. (10.41 kg.) per h.p. day; this corresponds to an output 
of i metric ton per .262 h.p. year. These later and better figures 
are the average of the first week's run after again starting up, 
and are attributed to the improved gas circulation, under the 
furnace roof which, according to Robertson, the inventors main- 
tain that the important point is to make this furnace last as 
long as possible, and in order to do this they consider it absolutely 
necessary to have the roof cool. Richards suggests (A.E.S., 
April, 1912) that the arch of the furnace hearth be protected 
by water-cooled plates, as is common with open-hearth practise. 
This, however, as has already been suggested, may decrease the 
efficiency too much. Lyon states that attempts were made at 
the Noble Electric Steel Co. in California to preserve the roof 
of the crucible hearth by water-cooled plates embedded in the 
brickwork, but these did not prove especially effective. Leffler 
writes at this time that they would gladly dispense with the 
artificial gas circulation if they could. As is elsewhere men- 
tioned, Leffier says that calcined limestone causes the burden 
to hang. Yet Noble, with his California furnace, says he only 
uses calcined limestone, and furthermore uses no artificial gas 
circulation.* In the last tests made at Trollhattan, the repairs 
and petty expenses cost about $1.60 per ton of pig iron produced. 
Part of these operating cost repairs are caused by the roof burning 
away. If half of the above amount were saved by the durability 
of the roof being increased, it would make, in a 2500 h.p. furnace, 
producing 25 tons daily, an annual saving of 350 X 25 X .80 = 
$7,000, enough to pay almost 9% on the investment. 

* The reason the Trollhattan furnace has gas circulation and the Noble 
furnace none, is because the former is operated as an arc furnace, but the 
latter as a resistance furnace. 


The cost of producing one ton of electric pig iron during 
the 5 months period ending in April, 1911, was estimated from 
the records and from a personal investigation given on the spot 
to be as follows: 

1.52 tons of ore, 67.1% at $2.68 $4.07 

.262 kw. year at $13.40 3.51 

85. kg. (187 Ib.) limestone at $1.61 per ton 14 

416 kg. (915 Ib.) charcoal at $12 4.99 

5.27 kg. electrodes consumed * at $67 36 

Labor 78 

Repairs and petty expenses 1 . 60 

Interest and sinking fund, 10% on $24,000 35 

Total $15.80 

The cost of producing one ton from hematite of 50% iron was 
$16.04. One of the Norwegian companies on the West coast, 
now (1912) constructing a plant for the smelting of 60% magne- 
site, estimated the cost per ton of iron, with electricity at $5.46 
a kw. year, at $11.25, using English coke at $5.63 per ton. 

For the Stassano furnace detailed cost figures are given by 
Osann in Stahl und Eisen, 1908, No. 19. They apply to the 
furnace described in Chapter VII for one-ton charges, making 
steel for castings from cold material. The figures are further 
based on the following special conditions. The furnace remains 
unused each night for three hours, and 24 hours on Sunday. 
During these times it is kept warm by electricity, the current 
being switched on for one-quarter of an hour, and off for three- 
quarters of an hour. Under this non-continuous operation the 
furnace gives 3.5 metric tons per day, or 840 metric tons per year 
of 240 working days. 

The furnace takes three men per shift, the average wage being 
given as $1.19. The lining costs $95.24, exclusive of the labor, 
when magnesite is used. It must be renewed every three weeks, 
that is, after a production of about 63 metric tons, and requires 
4 to 6 days for the renewal. 

* The total electrodes used per ton of iron produced was 10.28 kg., the 
difference being attributable to the stub ends, now no longer prevalent, with, 
the new screw type electrode. 


The construction cost of the furnace is given as $8,750. 
Under these conditions the following calculations are given 
per metric ton of fluid metal: 

Depreciation $ o . 992 

Cost of the charge : 

i metric ton scrap at $15.95 $15-952 

.02 metric ton mill scale at $4.047. ... .081 

.02 " lime at $2.857 .057 

.008 " " 12% ferro-silicon at 

$35-71 285 

.004 " " 8o%ferro-manganeseat 

$52-38 209 

.0008 " " aluminum at $357.00. . .285 


Cost of power : 

For melting 900 kw. hrs. at 1.0710 $9.643 

For heating during delays 1 .071 

Cost of furnace, lining, and repairs 2.619 

Labor 2 . 047 

Electrodes . 595 

Cooling water .095 

Total $33-941 

According to a more recent article (Neumann, Stahl und 
Eisen, 1910, p. 1066) it is possible to greatly reduce the cost 
of the lining when using dolomite for the hearth. At the same 
time through the use of a purer charge the power consumption 
for melting drops to 750 kw. hrs., and because of the correspond- 
ing less work with slags the furnace can last 70 to 100 heats. 
Definite figures for the lowering in costs brought about in this 
way are not known. 

Cost calculations for the Heroult furnace are similarly not 
known. On the other hand they have been published for the 
Girod furnace. The following are taken from Stahl und Eisen, 
1908, p. 1825, and apply to a 2-ton furnace. 


If a cold charge is worked, consisting of scrap, turnings, and 
pig iron, and completely refined to give steel of high value, the 
costs per metric ton are as follows: 

.1 ton lime, .1 ton ore, and additions of various 

alloys, from $ .714 to $ .809 

Electrodes 952 i . 190 

Labor 1.142 

Furnace maintenance, tools, etc 2 . 857 

1,000 kw. hrs., the cost depending on the price 

of power 

If melting, without further refining, is all that is necessary, 
that is to say a similar method of working to that recently car- 
ried out with the Stassano furnace at Bonn, then the following 
figures should be used: 

Lime, etc $o. 238 

Electrodes 762 

Labor 762 

Furnace maintenance, tools, etc 2 . 285 

Power consumption, 750 kw. hrs 

For the refining of a liquid steel charge taken from the con- 
verter or open hearth, the following figures are given: 

.04 tons lime, and additions $ .524 to $ .619 

Electrodes .381 

Labor .571 

Furnace maintenance, etc .952 

Power consumption about 300 kw. hrs 

In these tables depreciation and the loss in operation have 
not been taken into consideration. 

The latter is given by Borchers as 10 to 11%, who also says 
that the consumption of electrodes in the larger furnaces ought 
to amount to 0.012 to 0.015 metric tons per metric ton of steel 
with cold charges. This gives an electrode cost. of $0.571 to 
$0.762 per metric ton. In regard to the life of the furnace it is 
stated that with cold charges the walls last about 80, and the 
bottom about 120 heats, the roof stands 25 to 30 heats with small 
furnaces and 20 to 25 with large ones. 


The costs with the Heroult furnace will scarcely differ in an 
important degree from those of the Girod. As complete figures 
have not been given for the Heroult furnace the following partial 
results may be shown taken from Metallurgical and Chemical 
Engineering, 1910, p. 179. They apply to the 1 5-ton furnace 
at So. Chicago. Liquid Bessemer metal, of which the composition 
is not given, is refined in 12- to i4-ton charges, the final material 
containing 0.03% P. and 0.03% S. Power consumption 200 kw. 
hrs. per metric ton. The furnace roof of silica brick costs $60. 
It requires changing each Sunday.* With 12 heats a day, and 
13 metric tons per charge, this equals $0.0642. 

Hearth repairs, about $0.0642 

Lining costs, about 1284-! 

This does not take into consideration the costs per ton of door 
bricks, which must be replaced at certain times. The electrode 
consumption is given as 6.6 Ibs. per metric ton. Graphite 
electrodes are used, and the cost per ton of steel is about $1.504 

Neumann gives the loss with a cold charge as 6%, and 2.5 
to 3% with fluid charges. With the same kind of charge, how- 

* The roof problem has recently been the subject of careful study by 
FitzGerald (see A. I. E. E., June 25, 1912, transactions). A brick made of 
silicon carbide has been manufactured which it is believed will have a much 
longer life in the steel furnace than the silica brick now used. The brick is 
made by taking powdered or granular silicon carbide, mixing it with a suitable 
temporary binder, such as a solution of dextrine, molding and then heating 
in an electric furnace to the temperature at which silicon carbide is formed. 
Bricks made in this way have been used in the roof of an experimental steel 
furnace in one of these laboratories and then put to the severest test possible. 
The bottom of the furnace was purposely raised well above the normal level 
so as to bring the surface of the slag as close to the roof as possible, the actual 
distance in some experiments being only 10 in. (25.4 cm.). Then the furnace 
was worked at double the normal rate of generation of energy so that the heating 
of the roof was very intense, so much so that an ordinary silica roof would 
melt down rapidly and be completely destroyed in a single heat. Even under 
these very severe conditions the silicon carbide roof stood up perfectly. Ex- 
periments have also been made in other steel furnaces and these results con- 
firmed. The most serious objection to a roof of this kind is its relatively 
great cost, but if it lasts a sufficiently long time it is nevertheless economical. 

f Dolomite taken at $6.00 per metric ton. 

t This applies to electrodes of Acheson graphite, costing 50 cents per kg., 


ever, the same loss is to be expected in both the Heroult and 
Girod furnaces. 

A certain difference in the operating costs of the Heroult and 
all other electric furnaces arises from the fact that the former 
uses carbon for deoxidation instead of ferro-silicon. With normal 
heats, therefore, and deoxidation with carbon the Heroult fur- 
nace has to figure on a consumption of about 4 kg. ferro-silicon 
per metric ton; in the case of other furnaces, and the Heroult 
also, if deoxidation with carbon is not followed, on about 7 kg. 
If it is assumed that the Heroult uses 3 kg. petroleum coke, 
then the following figures are given for deoxidation and de- 
sulphurization : 

Heroult furnace, using 3 kg. petroleum coke at $1.90 $0.057 

Heroult and all other furnaces using 3 kg. ferro-silicon 

at 7.26c 216 

So that deoxidation and desulphurization by means of 
ferro-silicon alone is dearer than that by carbon and 
ferro-silicon by about 159 

As mentioned before, this has the advantage, however, that 
it does not influence the composition of the bath, and so is often 
used even in the Heroult furnace. 

There remains finally the operating costs of the Rochling- 
Rodenhauser furnaces. Such figures have been published by 
Wedding. The following apply to a 5-ton furnace working on 
fluid charges: 

Power consumption 230-280 kw. hrs. per metric ton 

Additions, about $o . 536 

Lining costs, using magnetite 595 

Wages 178 

Air for cooling transformers 050 

such as are used in this furnace. Up to within a short time it was impossible 
to construct satisfactory electrodes of carbon for the 1 5-ton Heroult furnace. 
This is emphasized by T. W. Robinson's discussion before the American Iron 
& Steel Institute, May, 1912, where he says: "Our necessities represented a 
requirement that the electrode manufacturers of America and Europe had 
not been called upon to meet, and it took much time and money before there 
was finally accomplished the 2O-inch round amorphous carbon electrode that 
is now being used" (in the 15-ton Heroult furnace). 


A 2-ton furnace using polyphase current gave the following 
costs, when scrap was worked up for making steel castings: 


I metric ton scrap $15 . 952 

5% loss 798 

.01 metric ton roll scale (22. Ib.) 040 

.035 " " lime (77. Ib.) 100 

.005 " " fluor-spar (15.7 Ib.) 074 

.01 " " sand (22. Ib.) 014 

.004 " " ferro-manganese . . . . ( 8.8 Ib.) 209 

Loss in ferro-alloys remaining behind 157 

Power consumption, about 900 kw. hrs., price varies. 

Lining and repair costs 636 

Labor 793 

Air for cooling transformers at 1.0710. per kw. hr 079 

These figures are given for a 2-ton furnace which, working 
with cold charges, allows a production of 6 to 8 tons per day. 
Apart from this it should be mentioned that the lining and 
repair costs when dolomite is used, and liquid charges, only 
amount to 0.238 to o.428c. with 3- to 8-ton furnaces. 

It may be mentioned again that all the cost figures given 
above are only exactly correct for certain predetermined local 
conditions. Care should, therefore, be taken in using them for 
comparison. The weight of material used ought to be shown, 
and the kind of charge and the final metal required have a great 

The consideration of the different factors affecting costs given 
in the first part of the book appear, therefore, to be very valuable, 
and this part may once again be referred to. 



UNTIL the invention of the steam-engine the operation of an 
iron and steel plant required the presence of a waterfall as the 
source of power for the hammers and blast. If, at the same time, 
sufficient ore beds and forests were in the neghborhood all the 
requirements were filled for the prosperity of the plant. The 
consumption of iron and steel tools was moderate, the plants 
could operate economically in a modest way and with small 
water-powers, for with the absence of railroads, etc., the products 
found a paying market in the immediate vicinity, and the bring- 
ing in of foreign goods was almost impossible. The few specially 
large German water-powers were not needed, and would not be 
used because the technical knowledge necessary was not sufficient- 
ly advanced. 

Conditions changed as the supply of charcoal began to de- 
crease and the consumption of iron and steel to increase, for the 
old plants with their associated water-powers and limited 
amounts of charcoal could only increase their production to a 
certain amount. The knowledge that ores could be smelted with 
coke, and the invention of the steam-engine, made it possible to 
use commercially the immense stores of energy lying dormant in 
the earth in the form of coal. Soon the plants deserted their old 
places near the waterfalls, and changed their locations to the 
coal-fields, where fuel and therefore power were present for 
application in unlimited amounts. 

Then succeeded the remarkable newer growth of the iron 
and steel industry with its attendant immense production. 

But the consumption of iron and steel constantly increases, 
coal begins to decrease in amount and become more expensive, 
and the industry will soon be forced, as in the time of our fathers, 
to look for a new and constant source of power. Electric energy 



is the first to come into consideration, since it is possible to pro- 
duce it from coal at a moderate cost. Also the railroads have 
brought the most remote countries into connection, and the 
enormous water-powers of foreign lands can be used as sources 
of cheap electric power. Is it to be wondered at that many 
technical men are working at the problem of the building of 
electric furnaces, or that this task should soon be solved economic- 
ally, when it is known that electric heating produces a higher 
furnace efficiency than heating with fuel? 

So we see efforts being made recently to build plants near 
the larger water-powers, as in the old days, in order to obtain 
electric power at the lowest cost, and to produce iron and steel 
from ore by electricity. 

Also in the industrial countries the electric furnace is gaining 
importance from day to day, for it is proving capable of pro- 
ducing higher quality steels equal to crucible steel, from impure 
raw material. 

It is the authors' wish that the production of iron and steel 
by electricity may receive such an impulse that the statements* 
in this little book will very soon be exceeded by the facts. 



The usual commercial process by which pig iron is produced 
is smelting in a blast furnace with fuel, flux and a blast of air. 
In the upper part, or shaft, of this furnace a continuous series 
of thermal and chemical reactions take place, which reduce the 
iron and prepare it for its final smelting in the hearth. These 
preliminary reactions could, if desired, be carried on in a special 
shaft into which ore is charged and subjected to the action of 
the hot furnace gases. 

In the lower part, or smelting zone, of the furnace the reduced 
and partially carburized iron is melted; the impurities of the 
ores and fuel are fluxed with the flux added for this purpose, and 
thereby converted into a liquid cinder, or slag. Besides these 
thermal effects, some chemical reactions occur which the temper- 
ature in the shaft was not sufficiently elevated to effect, such as 
the reduction of the oxides of silicon, manganese and phosphorus 
(the reduced elements being then absorbed by the iron), the 
conversion of iron sulphide in part to calicum sulphide, etc. 

First let us collect the data upon which to base a study of 
these reactions. Such data are given below; some of them have 
not been wholly confirmed experimentally, yet the estimated 
values are close enough to afford calculations of practical value: 

i kw. hr. = 864.5 cals. 
i kw. hr. = 1.34 h.p. hrs. 
i h.p. hr. = 0.746 kw. hrs. 
Spec. ht. of iron = 0.20. 
Spec. ht. of blast furnace slags = 0.30. 
Spec. ht. of ore = 0.20. 
Spec. ht. of CO = 0.243. 
Spec. ht. of C = 0.20. 

Spec. ht. of air according to its weight = 0.30. 
Latent heat of fusion of pig iron = 46 cals. 
Latent heat of slag = 30 cals. 





Test pieces of seamless drawn electric steel tubes (Rochling). 
normal tube, and test pieces made from it in the cold state. 








C. kg. CO kg. 

CO kg. 

C0 2 kg. 




FeO+CO = Fe+CO 2 
Fe 3 O 4 + 4C = 3Fe + 
4CO . 


















I,8 7 6 





- 803 
- 132 

- 941 

- 45 

- 74 






Fe 3 4 + 4 CO= 3 Fe 



p e2 O 3 +3C 2 Fe -f 


3 CO 



Fe,O s +3CO=2Fe + 
3CO 2 

MnO 2 +2C= Mn + 


Mn 3 O 4 + 4C = 3Mn 





MnO + C = Mn + 


SiO 2 +2C=Si+2CO 
P 2 O 6 + 5C=2P + 


To reduce 1000 kg. of iron from magnetite requires 1381 kg. 
of ore. For simplicity the ore may be considered as pure Fe 3 Oi 
without any earthy constituents which have to be slagged off. 
Reduction with pure carbon then takes place according to the 
following equation: 232 kg. Fe 3 O 4 + 48 kg. C = 168 kg. Fe -f- 
112 kg. CO. The CO therefore measures 4 X 22.4 = 89.6 cu. 
metres. For the production of a metric ton, 1000 kg. of pure 
iron 286 kg. of carbon are necessary and 533 cu. m. of carbon- 
monoxide are produced. 

In the blast furnace much larger amounts of carbon than 
these theoretical calculations call for are required, because 
carbon is depended upon not only to reduce the ore, but also- 
to furnish the heat required for the process. According to the 


equation Fe 3 O 4 + 4C = 3 Fe + 4 CO, the process can be 
carried out without the blast being used if the amount of heat 
is supplied which the table shows is necessary. The amount of 
gas produced by the reduction would be only about one-tenth of 
that produced in the blast furnace for the same weight of iron, 
for in the latter case the gas is diluted with a large nitrogen 

If magnetite is mixed with carbon in the proportion calculated 
above, and the mixture heated by ' electricity to the necessary 
reduction temperature as well as to the melting temperature of 
about 1300 C., reduction of the magnetite takes place readily. 
The following rough calculation gives the theoretical power 
consumption necessary for the production of i metric ton of 
iron in a condition fluid enough to be readily tapped, which is 
necessary in practise. 

1381 kgs. ore heated to 1300 C. = 1381 Xo.2 X 1300 = . . . 359,060 cals. 

286 kgs. carbon heated to 1300 .=286X0.2X1300= 74,360 " 

looo kgs. iron heated to reducing temperature 1000X941 = 941,000 " 

IOOO kgs. iron melted = 1000X46 = 46,000 " 

1,420,420 cals. 

This corresponds to - = 1643 kw. hrs. 


From this it is clear that the process requires much less 
carbon than the blast furnace if considerable electric energy is 
supplied. In the same manner a high-carbon iron can be 
produced if sufficient carbon be supplied not only to reduce 
the ore, but also to supply that which dissolves in the metal. 
A rough calculation for an iron with 3% carbon is given 
below : 

The carbon required is 286 + 30 = 316 kg., and 1030 kg. of 
pig iron is produced. 

1,381 kg. ore heated to 1300 C 359,o6o cals. 

316 kg. carbon heated to 1300 C 82,160 " 

1,000 kg. iron to the reducing temperature 941,000 " 

1,030 kg. iron melted 47,38o " 

1,429,600 cals. 


This corresponds to ~- - = 1653.7 kw. hrs. per 1030 kg. 


iron, which equals 1605.5 kw. nrs - P er metric ton. 

It should be remembered that the figures given for the coke 
consumption in the blast furnace take in all losses through 
cooling, radiation, etc., and in this respect the efficiency of the 
blast furnace is not bad. 

The power consumption given, on the other hand, is only the 
theoretical minimum, in operation it will be considerably higher, 
depending on the type of furnace used and its efficiency, etc. 
Also the figures for carbon consumption, are for chemically pure 
material, while in operation fuel containing ash has always to 
be figured on so that the minimum carbon consumption in the 
form of coke, charcoal or similar material is correspondingly 

The economical side of the smelting of ores by means of the 
carbon theoretically necessary for reduction and electrical energy 
to supply the heat for the thermal reactions requires that the 
saving in coke in the new process must be greater than the 
expense of the necessary electrical energy. 

As a result the process has prospective use only under con- 
ditions where ore and power are cheap and coke is dear, 
as in some parts of Canada, Italy, Norway, Sweden, Califor- 
nia, etc. 

The use of coke can be completely done away with and the 
iron separated from the ore electrolytically like aluminum, but 
the necessary power consumption is so extremely high that this 
method does not appear economical even for the future. Re- 
cently proposals have also been made to use iron pyrites as the 
raw material for smelting iron in the electric furnace. It is to 
be melted and air-blown through the bath until a consider- 
able amount of ferrous oxide has been formed; then the 
blast stopped, and the bath allowed to react according to the 

FeS + 2 FeO = 3 Fe + S0 2 

It is scarcely possible that the process will have a great 


The process given above of using just enough fuel to combine 
with the oxygen of the ore and electric heating of the ore-fuel 
mixture forms the basis of the many recent attempts to smelt 
iron ore by electro-thermal methods. 

It should be emphasized that, in electro-thermal processes, 
as the words themselves indicate, the electricity serves only 
as the source of heat which brings the charge to the temperature 
required for reduction and melting. Electrolytic processes 
where electricity is used both as a source of heat and as a reducing 
agent are less often employed because only direct current 
can be used. In regard to this, we may refer again to Part I. 

Electrical heating of the charge gives the great advantage 
that, because of the much lower fuel consumption, the influence 
of the latter on the charge and melted material can be regulated 
much better, and the operation can be carried out if desired at 
higher temperatures than used up to now in the blast furnace. 
This has a great metallurgical advantage for, as is well known, 
the "reaction ability" of all material increases considerably 
with increase in temperature. 

In general it is to be expected that in the smelting of the 
ordinary iron ores which contain more or less manganese, sulphur, 
phosphorus, silica, etc., the same reduction reactions will take 
place as are already known for the blast furnace process, etc., 
and that with electric smelting an iron of a certain determined 
purity and analysis will be obtained by regulating the furnace 
temperature and the slag. The iron will be very low in sulphur, 
for experience shows that the slagging off of the sulphur is 
favored by high temperatures, and with the electric furnace the 
temperature can be raised to any desired amount. Smelting in 
the electric furnace can also be carried on in such a way that, 
according to the amount of reducing material used, an iron can 
be produced of any desired carbon content, even practically free 
from carbon. However, it is a question whether it is preferable 
to produce right away a soft material, or to make a higher carbon 
product and suitably refine this later by special processes. Con- 
cerning this, local conditions alone can lead to a decision. 

In smelting lean iron ores, more electric energy is required, 


because the impurities have to be heated to the full temperature 
of the charge, and, furthermore, additional flux must be added 
to slag off these useless impurities, and the extra slag must also 
be heated to the temperature of the furnace. All this waste 
makes the process correspondingly more expensive. 

Raw spathic ore, brown iron ore, etc., must be calcined when 
smelted, which also requires electrical energy and correspondingly 
increases the cost. In conclusion: the electric production of 
iron, which is indeed an " infant industry," must be accomplished 
without the loss of an unnecessary kilowatt in order to successfully 
compete with the old economically working blast furnace. 

In general, therefore, at present usually high percentage 
iron ores, preferably magnetite and red hematite, are smelted 
electrically. If it happens that poorer ores have to be used, then 
they must be previously carefully prepared and concentrated. 
During this concentration it is well to remove as completely 
as possible any pyrites, apatite, etc., which may be present, 
and thereby help in the production of a highly valuable iron of 
great purity similar to Swedish or Styrian, which will be suitable 
for the production of high quality steels. The fuel must also 
be as low as possible in ash, so that the slag volume is not in- 
creased too much. The size of the material is of secondary 
importance for suitable reduction, but very fine materials are 
not willingly used exclusively because of the difficulty of removing 
the gases produced in reduction. 

In the first tests carbon and ore were intimately mixed, 
pressed together with tar and used in the form of briquettes. 
This briquetting is unnecessary and can be more readily rejected 
as it is costly, for in those countries where electric smelting is 
commercially possible because of dear coke the price of tar is 
also correspondingly high. 

Electric smelting of iron ore can be carried on in electric 
steel-making furnaces. The mixture for reduction will either 
be charged altogether, or else added little by little, depending 
on the type of furnace. If a pool of liquid pig iron has formed 
on the hearth, then the reduction of the ore mixture will progress 
more quickly, for the carbon of the liquid metal takes an energetic 


part in the reduction. The fluid pig iron will then have to 
be recarburized to the required amount by the carbon of the 

In regard to the necessary power consumption, that type of 
furnace will work most favorably with which the radiation loss 
is the smallest. 

The Stassano furnace, to the construction of which the first 
part of this book is devoted, heats the mixture by radiation, for 
the arcs are outside of the material to be heated. But, as the 
arcs radiate heat in all directions, and only that much which 
radiates downwards is used economically, it is to be expected 
that the efficiency of this furnace will be proportionately low. 
On the other hand, the electrode consumption will not be very 
high for the electrodes are not in contact with the charge, and so 
will not be attacked by the iron ore. 

i. Smelting of Ore in the Stassano Furnace. (The charge 
heated by radiation from the arc.) Neumann and Goldschmidt 
have published results of the following smelting test (Stahl und 
Eisen, 1904, pp. 687, 885). The analyses of the materials used 


Fe 2 O 3 93-02% P 056% 

MnO 62 CaO+MgO 5 

SiO 2 3-79 H 2 1.72 

S 058 


CaO 51-21% Fe 2 O 3 50% 

MgO 3.11 SiO 2 90 

A1 2 O 3 50 CO 2 43-30 


Carbon 90.42% Ash 3-88% 

Water 5.70 


Carbon 59 . 20% Hydrocarbons ... 40 . 50% 

Ash 27 

Briquettes were made from a mixture of 1000 kg. ore, 125 kg. 
lime, 1 60 kg. charcoal, 120 kg. pitch (charcoal and pitch together 


containing 230 kg. carbon). These briquettes constituted the 
charge. According to Stassano the heat requirements per 100 
kg. ore in the charge are calculated as follows, the data below 
being chosen from his tables. 

Decomposition of the oxide of iron - = 111552.000 

Vaporizing the moisture in the ore and charcoal 

(1.72 +1.21)637 = 1866.41 

Calcining the flux 12.5 X 425 = 53 I2 -5 

Heating the C0 2 to 500 C. 5 ' 429 X - l6 X 5 = 987.09 


Heating the CO produced - X .0068 X 500 = 5921.667 


Melting the iron produced 65 X 350 = 22775.2 

Melting the slag produced 13.89 X 600 = 8334.0 

Produced from the burning of C to CO 20.9 

X 2175 45457.500 

Leaving 111853.927 

These 111,853.927 calories correspond to 129,386 kw. hr. 
From 100 kg. ore 65.114 kg. of iron will be reduced, so that the 
power required per metric ton of iron is 1987.6 kw. hrs. This 
power requirement is, however, only calculated theoretically, 
and figures concerning the real power consumption have not 
been published; however, as shown above, the radiation loss 
with the Stassano furnace must be considerable. 

An idea of the amount of this radiation loss is obtained from 
a further test published by Goldschmidt in which the power 
consumption is given. In this test 70.25 kg. of the same 
briquettes used in Test No. i were smelted in a 100 h.p. furnace. 
The output was 30.8 kg. iron with a power consumption of 97.2 
kw. hrs. = 132.2 h.p. hrs. The theoretical power consumption 
for the charge may be calculated on the basis of the analyses 
given above as follows: 


For the reduction of the iron contained in the final product 
were necessary 3 727 ' 312 x 192 = 52730.262 

For the reduction of the manganese in the final 
product were necessary ^- x 94-6 = 48-7 I 9 

For melting the metal 30.8 X 350 = 10780.00 

For melting the slag 6.3 X 600 = 3780.000 

For heating and vaporizing the moisture 

1.316 X 637 = 838.292 

For calcining the lime 6.25 X 475 = 2968.750 

For superheating the steam to 500, 

1,316 X 400 X .48 = 252.672 
For superheating the CO 2 to 500 C, 

2.714 X 500 X .016 

= 493-554 


For superheating the hydrocarbons to 500 
2.43 X 5 X .27 = 328.05 

For superheating the CO produced 

(3 X 30727 X3i2 + 28.336) 

( 112 55 ) 

X 500 X .0068 = 2800.131 

Total 75020.330 

From the combustion of the C to CO were produced 
9.883 X 2175 = 21495.525 

Leaving 53524.805 

As the whole charge consumed 97.2 kw. hrs. = 84012.072 cals., 
the heat efficiency was: 

53524.805 X IPO _ 


The power consumed per metric ton of iron reduced from its 
ore is shown to be 3123 kw. hrs. Unfortunately the analysis 


of the metal produced is not given, nor the length of time of 
the test. The demonstrated efficiency of 61.33% * s n t very 
difficult from the calculated figure. It must be admitted 
that not only was this a test melt, but that several of 
the figures calculated gave accidentally very favorable results. 
In operation the efficiency would undoubtedly be much smaller, 
for the careful supervision possible with a small test would be 

Because of the great heat radiation in the furnace which 
principally attacks the roof, the life of the roof must be small, 
and the economical carrying on of the process depends in the 
first place on the durability of the furnace. In order to make 
the roof somewhat more durable, either the whole or at least that 
part attacked the most must be protected by water cooling. 
This water cooling, however, apart from its complexity will bring 
about important heat losses, the amount of which will be gone 
into further in another place. 

The power consumption per metric ton of iron is seen to be 
high as was to be expected. Theoretically it is 1643 kw. nrs - or 
2680 with a furnace efficiency of 61.33%, compared with a 
proved figure of 3123. This increase in practise of 443 kw. hrs. 
is due to the use of ore which is not theoretically pure, and the 
consequent melting of the slag produced, the burning of the 
lime, vaporizing the water, etc., a proof that only ores as pure 
as possible should be smelted. With the smelting of more im- 
pure ores the power consumption would naturally be con- 
siderably higher yet. This high power consumption is due to 
the great radiation loss of this type of furnace, and can therefore 
scarcely be lessened. Further disadvantages are that no con- 
tinuous operation is possible, and only small heats can be pro- 
duced. From this it is evident that furnace types in which, 
like the Stassano, the charge is only heated by radiation can not 
be considered in the economical smelting of ore. 

2. Ore Smelting in Electric Hearth Furnaces. (Electrodes 
introduced into the charge.) 

Theoretically these furnaces should work well because the 
charge so nearly surrounds the arc that the heat radiated is 


completely absorbed. In operation, however, such a total 
absorption is impossible, the charge can only surround the arc 
to a limited extent, and the temperature is so high that radiation 
through the charge to the walls of the furnace is unavoidable. 
With such furnaces the lining and the special roofs, if such are 
present, are particularly strongly attacked by the "stagnant 
heat," so that it is impossible to maintain continuous operation. 
Also the electrode consumption will be high, for the electrodes 
are in contact with the ore mixture and will be attacked. 

Many ore- smelting tests have been carried out with different 
types of furnace in recent years in order to remedy the trouble re- 
sultant upon the attack on the furnace walls, but with uncertain 
results. In every case the power consumption has been much 
more favorable than was expected, so that in this respect the 
question of electric smelting of ore would have been long since set- 
tled if a furnace construction had been found more suitable for 
continuous operation. The most recent tests of this kind have 
been carried out by Messrs. Gr on wall, Lindblad & Stalhane, 
the latest test furnace being shown by figs. 124, 125 and 126, 
invented by the same men. One metric ton of white iron was 
produced in 1909 with 0.25 h.p. years equals 2190 h.p., that is 

2 1 GO 

" =1622 kw. hrs., a result that closely approaches the 


theoretical minimum, and is to be explained perhaps by the very 
pure ore smelted. Further tests made with the Gronwall, Lind- 
blad, Stalhane furnace are given elsewhere in Chapter XIII, under 
"Operating Costs," and under B, " Electro-metallurgy of Iron." 


The efficiency of this furnace will not be bad for smelting 
ore, notwithstanding that the charge is only heated by the heat 
of the molten bath, because the bath is covered with cold charge 
and the radiation from the lower part of the furnace can be kept 
low by means of suitable brickwork, etc. 

Above everything else, however, because of the cooled upper 
surface of the bath due to the covering of the charge, and the heat- 


ing from within, the roof will show great durability, which is a very 
important point, if there is to be continuous operation. Those 
furnaces which work with electrodes and have a roof are com- 
pelled to use extensive water cooling, sometimes in order to 
increase the life of the furnace. Through the avoidance of water 
cooling, a source of considerable loss of heat is avoided, so that 
the induction furnace is worthy of serious investigation for the 
smelting of ore. 

Many smelting tests have been carried out in the Rochling- 
Rodenhauser furnace, and some reports of them may be given, 
for up to the present scarcely any results of ore smelting in the 
induction furnace have been published. High sulphur magnetite 
in a very fine state of division was used and high sulphur coke 
breeze, in order to produce a pig iron with 2.6 to 3.0% carbon, 
and as low in sulphur as possible. Although it was assumed 
that a greater part of the sulphur would pass away as gas due to 
the following reaction : 

FeS + 2 FeO = 3 Fe + SO 2 , 

yet, by way of precaution, the theoretical amount of lime neces- 
sary to combine with the sulphur as sulphide of calcium was 
added to the charge, together with that necessary for the acid 
gangue, etc. The amount of slag produced in this way was not 
needlessly increased, although the CaS produced requires a 
large amount of slag for solution if it is hoped to produce a 
sufficiently good desulphurization in this way. 

Analysis of the ore: 

Fe 3 4 96.38% = 69.79% Fe/ 

FeS 2 


= 0.66% Fe) 

Mn 3 4 


= o.i8%Mn. 

Si0 2 


P 2 5 


= 0.02% P. 




1. 21 


Oxygen combined with Fe & P = 26.62%. 
Total sulphur in the ore = 0.75 





Carbon 87 .48% 

Sulphur i .068% 

Ash 10.4% 

The principal constituents of the coke ash were: 

Si0 2 40.6 % 

CaO 5.6 % 

Fe ii. 6 % 

A1 2 3 25.40% 

Oxygen combined with iron = 5.0%. 
The carbon-monoxide produced passes away unused. 
Chemical Balance Sheet. 100 kg. ore (i/io of a metric 
ton) = 70.45 kg. iron. This requires: 

(a) For reduction and combination of the 26.62 kg. oxygen, 

C + = CO 
i60 + 12 C = 28 CO 

26.62 X 12 


= 19.97 k g- 

, , 26.62 X 28 , , --. 
and produce - = 46.58 kg. CO 

(b) 70.45 kg. Fe carburized to 3% require 

7045^3 = 2-j8 kg _ c 

(c) 0.75 kg. S combined with CaO to form CaS require: 

S + CaO + C = CaS + CO 

32 + 56 + 12 = 72 + 28 

0.75 kg. S + 1.28 kg. CaO + 0.28 kg. C 
= 1.70 kg. CaS + 0.70 kg. CO 

The total amount of carbon is therefore 19.97 + 2.18 + 0.28 
= 22.43 

,. 22.43 X 100 

corresponding to - - = 25.65 kg. coke breeze. 



As, however, the ash of the coke also requires a small amount of 
carbon for the reduction of its metallic oxides, the calculations 
should be made with 25.80 kg. of coke breeze. This 25.80 coke 
breeze contains 0.258 X 1.068 = 0.28 kg. sulphur which must 
be slagged off as CaS. 

0.28 kg. S + 0.49 kg. CaO + kg. C = 0.63 kg. CaS + 
0.24 kg. CO. 

25.80 kg. coke breeze contains 



- - =0.13 kg. O, in the form of oxide of iron. 

16 O + 12 C = 28 CO or 
0.13 O + C = 0.23 CO 
100 kg. ore therefore require 
22.43 + + = 22.63 kg. carbon or 

= 25.87 kg. (57.03 Ibs.) coke breeze. 



From the ore From the coke ash 

0.2587 X 10.4 = 2.69 kg. 
SiO 2 0.60 kg. Si0 2 i. 09 kg. 

CaO kg. CaO 0.15 kg. 

MgO i. 2 1 kg. A1 2 O 3 0.69 kg. 

1.91 kg. 2.69 kg. 

from this must be taken 0.166 X 10.4 X 0.2587 = 0.44 kg. Fe 3 4 , 
leaving 2.69 0.44 = 2.25 kg. slag. 

CaS produced 1.70 + 0.63 = 2.33kg. 

Lime addition for combining with the sulphur 1.28 + 0.49 = 1.77 
Lime addition for slag ................................... 1.41 

Total lime addition ............................... 3.18 


The theoretical total amount of slag is 1.91 + 2.25 + 2.33 + 
3.18 = 9.67. 

In calculating the amount of carbon necessary for reduction, 
it must be remembered that before the beginning of the test 
the furnace was filled with 1000 kg. of refined Basic Bessemer 
metal, which latter had to be recarburized to the required 
amount. After this the following mixture was charged: 

597.5 kg. ore + 183.5 kg. c ke breeze + 19.0 kg. lime = 800 kg., 

compared with the theoretical amount which does not consider 
the recarburization of the refined Bessemer metal: 

597.5 kg. ore + 154.6 kg. coke breeze + 19.0 kg. lime. 

The Bessemer metal had a temperature of 1650 C. The ore 
was charged as uniformly as possible, and in comparatively large 
amounts. Care was taken that the bath was always covered 
with the mixture in order to keep the radiation loss as low as 
possible; a method of working that, in general, was not hard to 

The slag produced during the tests was only removed once, 
and the exact amount was obtained. As the furnace used for 
the tests was mounted on a scale, the weight of the Bessemer 
metal charged and the finished material were also obtained 
exactly. The smelting of the 800 kg. of charge required 
1030 kw. hrs. 


(a) Output of Slag. 

99 kg. slag with 9.12% FeO = 7.09% Fe and 2.60% S. 
Theoretically the slag should contain: 

1. From the charge 9.67 X 5.975 = 57.78kg.. .57.78 kg. slag. 

2. Lime for slagging (3.18 - 1.77) = (1.41 X 

5.975) = 8.42 kg. slag. 

3. Excess of coke breeze = 3.00 kg. slag. 

4. Slag remaining in furnace from previous 

heat = 20.8 kg. slag. 

Weight of slag = 90.00 kg. slag. 


The weight of this slag is increased by its iron contents 

90 X 100 

- = 99 kg. which contains 
loo - 9.12 

7.09 X 99 , 

- = 7.02 kg. iron. 

(b) Output of Iron. 

looo kg. of Basic Bessemer was charged containing: 

C .06% P .093% S .073%, 

1427 kg. electric pig iron were tapped with 2.64% C, 0.02% Si, 
-73% P> 0.076% S, of which 1427 1000 = 47 kg. were 
produced from the mixture. The theoretical amount is 

(1) From the ore 5.975 X 0.7045 = 420.94 

(2) From the coke ash, 

1.835 X IJ -6 X 10.4 2.20 

= kg. iron = - 
100 423-14 

*7 O2 

From this kg. iron = - ^ went to the slag. 

This weight when calculated to electric pig iron equals 


= 427 kg. 


The loss of metal in the slag is therefore- = 1.6%. 


(a) The Carbon. 

1 ,000 kg. Basic Bessemer metal with 0.06% C. carbur- 

ized to 2.64% C. require (2.64 .06) 1,000 = . . . 25.8 kg. C. 

427 kg. electric pig iron contain 427 416.12 = 10.88 

The reduction process requires 5.975 X 19-97 = H9-3 2 ' 

The formation of CaS requires (0.28 + 5.975 =.. 2.27 ' 
The reduction pf the iron oxide in the coke ash requires X 5-975 = 

I59-25 kg. C. 

As 87.48 kg. C = 100 kg. coke breeze, this 159.25 kg. C 

182.0 kg. coke breeze. 


(b) The Sulphur. 
Brought in: 

597-5 kg. ore at 0.75% S= 4-4$ kg. S 

183.5 kg. coke breeze at 1.068 S= I -9 6 ' 

20. 8 kg. slag held back containing 1.0% S= . . . 0.21 ' 

6. 65 kg. S 

Taken out: 

99 kg. slag at 2.6% S = 2.57kg. S 

427 kg. electric pig iron at 0.076% S= 0.32 ' 

2 . 89 kg. S 

Therefore 6.65 - 2.89 = 3.76 kg. S or 56% of the total 
sulphur was gasified. 

(c) Phosphorus. 

1000 kg. Basic Bessemer metal at 0.093% P = . . . . 0.93 kg. P 
597.5 kg. ore at 0.02% P= 0.12 " 

P brought in 1 . 05 kg. P 

1427 kg. electric pig iron at 0.073 nee d = 1 . 05 kg. P 

(d) The Furnace Gases. 

12 C + i60 = 28 CO 

Therefore (119.32 + 2.27 + 0.98) = 

122.57 k g- C + 163.43 kg. O = 286.0 kg. CO 

183.5 182.0 = 1.5 excess coke cor- 
responds to 1.31 C. This was 
charged in excess, burned with 
air gives 3.05 kg. CO 

The burning takes place with 1.74 kg. 

O, that bring in 6.54 kg. CO 

The total gas made is 295. 59 kg. 

The furnace was held at 1300 C. during the test, and the 
iron tapped at the same temperature. The mixture for reduction, 
therefore, had to be first heated to this temperature after charg- 


z. Heat Expended. 

1. 597.5kg. ore require 597.5X1300X0.20= 155,350 cals. 

2. 183.5 kg. graphite, 183.5X1300X0.23= 48,426 " 

3. 19. o kg. lime require 19X1300X0.21 = 5,187 " 

4. 416.12 kg. iron reduced from FesO 4 require 1648 

X4I6.I2 = 685,765 " 

5- 5,975Xo.O2 kg. P to be reduced from P 2 O 5 re- 
quire 0.12X5966= 716 " 

6. 427 kg. pig iron require for melting 427 X46 = . . 19,642 " 

7. 90 20.8=69.2 kg. slag require for melting 69.2 

X30 = 2,076 " 

8. 2.89kg. S changed into CaS require 2. 89X1093= 3,159 " 

920,321 cals. 

2. Heat Brought in. 

1,000 kg. Basic Bessemer metal cooled from 1650 

to 1300 bring in 1000X350X0.2 = 70,000 cals. 

Burning 3.76 kg. S to SO 2 = 8,347 " 

l8 3-5 kg. coke breeze = 160.53 kg. C. For carbur- 
izingthepig iron 26.40 + 10.88 =37.28 required. 
The remainder, 123.25 kg. burned to CO bring 
in 123.25X2473= 304,797 " 

1030 kw. hrs. used in the test 1030X864.5 = 890,435 " 

Heat brought in = 1,273,579 cals. 

Therefore the efficiency of the furnace equals 
920321 X IPO 


In determining the energy necessary to produce i metric ten 
pig -iron it must be remembered that the basic Bessemer metal 

charged at 1650 brings in |^ = 81 kw. hrs., for the finished 

material was tapped at 1300 C. Therefore 1030 + 81 = mi 
kw. hrs. were required to produce 427 kg. electric pig iron, which 

equals for the metric ton = 2602 kw. hrs. 


The following important points were established by means 
of the test. 

(i) The efficiency of the furnace is good, as was to be expect- 
ed. It may be still further increased if the mixture for reduction 
were charged by machinery and not by hand, so that the frequent 


opening of the working doors and the unavoidable heat losses 
could be avoided. 

(2) The reduction of the ore takes place satisfactorily even 
with the use of dense fuel, chemically inactive, such as coke 
breeze with a high content of ash. The amount of reduction 
material necessary closely approaches the theoretical, due to the 
reducing atmosphere of the electric furnace. 

(3) The phosphorus in the charge goes entirely into the iron. 

(4) The sulphur in the charge is lowered more than half, due 
to the reactions between the oxides and sulphides of iron, so that 
lime additions to unite with the sulphur are probably unneces- 

Test 2. In order to increase the efficiency of the furnace ef- 
forts were made to lower the radiation from the upper surface of 
the bath by causing the charge to project still deeper into the iron 
bath in the hearth. A suitable way appeared to be the smelting 
of the charge in the form of briquettes. The briquettes were 
made of the same mixture as used in Test No. i plus 8% of steel- 
works tar. The whole was ground in a Chili mill, and pressed 
in an ordinary dolomite press. The briquettes were burned a 
little before being used. An interesting point is that these 
partially burned briquettes showed 0.35% reduced metallic iron. 

A lowering in the power consumption with the use of these 
briquettes could not be proved, nor any increased smelting 
efficiency of the furnace compared with Test No. i. The pig 
iron produced had a low sulphur content, and the chemical 
balance showed that a greater part had been gasified as SO*. 
The same amount of ore was smelted as in Test No. i. 

Test 3. As both tests showed that over half the total sulphur 
was gasified, and the iron was sufficiently low in sulphur, further 
tests were made on a mixture of ore and fuel without special lime 
additions. It was thought that because of the smaller slag 
volume, the power consumption would be lower, and that at the 
same time the iron would be sufficiently low in sulphur. 

After Test 2 had shown the lack of efficiency of briquetting, 
the ore was used fine as it was taken from concentrating, and the 
coke breeze of the usual size. The results of the test were good. 


A low sulphur white iron was produced, and somewhat fewer 
kw. hrs. per metric ton were necessary than with Test i, and the 
furnace efficiency was somewhat higher; that is, the smelting 
time was somewhat shorter than with Tests i and 2. 

Test 4. This test was to show whether an addition of 
granulated iron to the mixture would shorten the time of melting, 
and give a saving in the energy consumed. It was really a 
carrying out of the so-called Lash process, which consists of 
using a mixture of ore, carbon, and slag-producing material with 
finely divided pig iron, the charge being kept loose and porous 
with sawdust. The reduction of the ore is helped by the carbide 
present in the pig iron. An example of a Lash mixture is as 

Iron ore 54% 

Cast-iron turnings or granulated cast-iron 27 

Sawdust 4 ' 

Limestone 4 

Tar 3 

Coke 8 


From what has been said before, it is to be expected that 
ore reduction by the Lash process would give no advantage, for 
in the induction furnace there is present a permanent bath of 
metal, and therefore with the ordinary ore mixture the known 
good reactions in the Lash process must take place anyway. In 
melting a metric ton of pig iron by the Lash process, the power 
consumption will be rather bad because the iron enclosed in the 
charge has to be melted electrically. 

The reduction mixture was charged in exactly the same way 
as described by Lash. The result of the test, however, gave 
neither a shorter melting time nor a lower power consumption 
per metric ton of pig iron from ore. 




The smelting of iron ore in the electric hearth furnace, which 
is so simple experimentally, depends on two important factors 
before it can be carried out commercially. One of them is the 
power consumption, the other the durability of the furnace lining, 
that is, the costs for repairs per metric ton of iron produced. 

The durability of the lining requires that the highest tem- 
peratures, such as those of the arc, must be avoided because the 
drop in temperature is too great for it to be taken up by the 

This question of smelting ore in the electric hearth furnace is 
therefore only to be solved by a type of furnace that does not 
work continuously at the highest temperatures, and with which 
the excess heat which attacks the lining can be carried off. In 
this case the lining costs will be very small, but a somewhat 
higher power consumption must be counted on. 

From the discussion above the only furnace of this type at 
present is the induction furnace, and the tests show that on the 
one hand the furnace lining allows continuous operation, and on 
the other that the power consumption is within such limits that, 
under certain conditions, successful competition with the blast 
furnace is permissible. Such conditions are first, that there are 
no special requirements in regard to the physical properties of 
the ore and fuel. Even very fine raw materials can be smelted, 
but the best are of small grain size. 

This factor becomes more important from day to day, for 
conditions continually press towards the mining of poorer grade 
ores and magnetic concentration, so that high percentage 
concentrates, small in size, are coming on the market. If these 
concentrates are to be smelted in the blast furnace, they must be 
first agglomerated or briquetted, a process that even without a 
binding agent, that is, using high pressure alone, or say sinter- 
ing, is an additional expense, for in this case a preroasting 
cannot be avoided. 

Also in smelting ore in the electric hearth furnace a small 


sized material can be used for reduction with at least equal 
success to one of moderate size, which means that small waste 
fuel of any kind is available that up to now has been valueless. 
Even these two points are so important that under certain con- 
ditions they will allow the electric hearth furnace to work more 
economically than the blast furnace. Also in regard to the 
purity of the ores, especially the sulphur, the electric hearth 
furnace has great possibilities because of the considerable 
volatilization that takes place. High sulphur materials can be 
smelted, therefore, with acid slags and without the lime additions 
that are absolutely necessary in the blast furnace. Only so 
much basic flux need be charged as is necessary to give a 
liquid slag. 

The concentration of the ores will therefore not have to be 
carried so far, especially when ores with an acid and basic gangue 
are to be used at the same time, for by suitable mixing a self- 
fluxing charge can be obtained. This allows the conclusion to be 
drawn, that under certain conditions the poorer iron ores can be 
smelted in the electric hearth furnace without previous prepara- 
tion, especially if the gangue forms a flux, so that the iron output 
of the charge is not lowered by the addition of fluxes. 

A further important advantage of smelting ore in the electric 
hearth furnace is that the harder steels can be produced direct. 
It is not favorable to immediately make a soft steel, for the iron 
bath is first carburized by the reducing material, so that at the 
end of the heat ore alone must be added in order to remove this 

In the next section of the book it is explained how this process 
is comparatively expensive. Still steel with about 1.5 to 1.8% 
carbon can be produced direct, and if high in sulphur can be 
desulphurized at little cost; while, at the same time, if high in 
phosphorus it can be dephosphorized without removing the 
carbon, both by means of processes given in more detail in the 
next chapter. 

The carbon consumption in the electric hearth furnace is as 
good as possible when the carbon is only burned to carbon-mon- 
oxide. Troubles that are always more or less unavoidable in 


the blast furnace disappear altogether, as also the production of 
the valueless "transition iron/' when the furnace is changed 
from one kind of iron to another. Add to this the simpler 
operation, the avoidance of water cooling, the possibility of reg- 
ulating at will the temperature of the metal tapped, and no 
electrode consumption, are some of the results. All these are 
points that, under certain conditions, allow the electric hearth 
furnace to successfully compete with the shaft furnace for smelt- 
ing ore. 



The experiments made in the electric hearth furnace make 
one desirous of studying more economical methods of smelt- 
ing. The disadvantages of the electric hearth furnace are 

1. Low melting efficiency of the furnace during operation. 

2. Large power consumption per ton of iron produced. 

3. Frequently too high a consumption of reducing material. 
The reason for the low furnace efficiency is that the mixture 

for reduction is charged cold so that it has to be heated electrically 
to the necessary temperature. As the radiation loss increases 
with the smelting time per ton, it follows that a shortening of 
the smelting time would give a better efficiency, and this requires 
the charging of heated material. This preheating must naturally 
be brought about without increased consumption of electric or 
other energy if possible, and the hot waste reduction gases are 
available without extra cost. They are most suitably used by 
charging the mixture high in the furnace so that the gases have 
to pass through it, giving up their heat. This necessitates 
arranging a shaft on the hearth furnace. 

The carbon-monoxide produced in the hearth would not only 
have a thermal effect but also a chemical one, that is, the ore 
would be partly reduced, so that the furnace then has only to 
melt the iron in the mixture of iron and ore. In other words 
only the remainder of the iron ore has to be reduced, and the 
furnace is released from some of its work. 


If the reduction gases are used only to preheat the charge, 
then the following rough calculation gives the ad vantage obtained: 
533 cu. m. (1880 cu. ft.) of gases are produced per metric ton of 
iron, and the same may be used for preheating the charge and 
be cooled down to 200 C. There will be obtained therefore: 

533 X 1 100 X 0.24 = 140712 cals., corresponding to / = 


162.8 kw. hrs. In producing a pig iron with 3% carbon there 

u u 162800 m , . , 

would be a saving m energy of - -- = 10.14%, which means 

that the metric ton of pig iron will be smelted with 1605.5 
162.8 = 1422.7 kw. hrs. 

On the other hand if the waste gases are used only for pre- 
liminary reduction of the ore, then the following rough calculations 
are obtained for the limiting case that the CO is all changed to 
C0 2 . According to the equation Fe 3 4 + 2 C = Fe 3 + 2 CO* 
the metric ton of iron would only require 143 kg. of carbon for 
reduction. Also, according to the equations: 

Fe 3 4 + 4 C = 3 Fe + 4 CO 
Fe 3 4 + 4 CO = 3 Fe + 4 C0 2 

only - - = 493000 cals. 

would be necessary. The total carbon required for the produc- 
tion of a 3% carbon pig iron will be 143 + 30 .= 173 kg. for 
1030 kg. metal, and the following heat balance is obtained for 
this most favorable case. 

1383 kg. ore heated to 1300 C. = 1381X0.2 

X 1300. ............................... =359,o6o cals. 

173 kg. C. heated to 1300 C. = 173X0.2X1300 = 44,980 " 

1000 kg. iron heated to reducing temperature .. =493,000 " 

looo kg. iron heated to melting temperature ... = 47,880 " 

944,920 cals. 

This corresponds to - = 1093 kw. hrs. per 1030 kg. pig 

iron, or 1061 kw. hrs. per metric ton. It is here assumed that 


the C0 2 leaves the furnace at 1300 C., and if the excess heat 
of the CO 2 were further used to preheat the charge, and the gas 
allowed to escape at 200 C., then the power required would be 
lowered as follows: 

From the equation Fe 3 O4 + 2 C = Fe 3 + 2 C0 2 
168 kg. Fe produce 2 X 22.4 = 44.8 cu. m. CO 2 , or 268 cu. m. 
per metric ton of iron. If the heat from 1300 to 200 is used 
for preheating then there is obtained 268 X 0.24 X noo = 70752 

cals., corresponding to "^" = 81-8 kw. ^ rs - -^ n this case > 


therefore, 1061 81.8 = 979.2 kw. hrs. are necessary to pro- 
duce i metric ton of pig iron. 

From this it may be seen that the use of the furnace gases 
for reducing the ore brings about a considerable lowering in the 
power required, just as well as their use for preheating alone. 
By utilizing these gases as much as possible, the electric furnace 
is relieved a great deal and the smelting time is considerably 
shortened. The idea of using the reduction gases is therefore 
justified particularly as, at the same time, there is obtained 
a desirable and much lower consumption of reducing 

As is well known, however, carbon-monoxide can only 
be used up to a certain limited amount for the reduction 
of ore because the mixture of CO and C0 2 produced has no 
more reducing influence when a certain percentage of CO 2 
is present. In the electric shaft furnace, therefore, one has 
to figure on a waste gas that consists largely of CO, 
and it is apparent that the carbon necessary for reduction 
will increase with increasing percentage of CO in the waste 

In smelting magnetite the carbon necessary per metric ton 
of pig iron with 3% carbon, when the percentage by volume of 
CO 2 in the waste gases is known, is calculated by the formula 

286 (IOQ - 3) + 30 
100 + C0 2 % 

Jn this way the following table has been prepared: 


Per cent. COz in Kg. carbon required per 
waste gases metric ton pig iron of 


ioo 138 

90 146 

8 154 

70 163 

60 173 

50 184 

40 198 

30 213 

20 231 

10 252 

o 277 

The principle of an addition of a shaft is naturally possible 
with any electric hearth furnace that has a fairly large hearth, 
and is the easiest in the case of the arc furnaces, for these always 
have a comparatively large hearth. The Stassano furnace forms 
an exception, for here the charge is heated by radiation alone and 
only the heat below the arc is used. Also the induction furnace 
can be built so that it is easy to add a shaft, and further as the 
depth of bath in the induction furnace can be fixed at any 
desired amount a shaft about 3 m. (10 ft.) high or over is per- 
missible, which is completely sufficient because of the small 
amount of reduction gases produced and their slow passage 
through the shaft. 

In principle, reduction with gaseous fuel is always preferable 
to solid fuel, for the latter only reduces the outer layers of the 
ore. Because of this the use of a gaseous reducing agent should 
shorten the time of operation and increase the efficiency of the 
furnace, for the reasons already given. At first it was feared 
that, with the use of a shaft, the heat would be immediately 
carried upward from the metal bath and the operation of the 
furnace thereby made more difficult. These fears, however, 
were shown to be groundless because preheating helped the 
furnace so that the same condition was obtained as before, but 
in a shorter time. It will be shown that the carrying away of 
heat from the hearth to the shaft only takes place slowly, and 
that in arc furnaces the heat must be artificially removed from 
the lower part of the furnace. 



In addition to the economic advantages of the electric shaft 
furnace compared with the hearth furnace, the disadvantages 
should not be overlooked. They are: 

1. No very fine material can be smelted, but only pieces that 
are not too large, nor on the other hand ore smaller than a hazel- 

2. In smelting there is no removal of sulphur, therefore with 
ores, etc., rich in sulphur there must be added the necessary 
amount of fluxes to slag off the sulphur. 

3. The slag must be tapped as a thin liquid, so that for this 
reason fluxes also must be added, which decreases the output 
from the charge. Therefore at present only high percentage 
ores can be used. 

4. Only iron with considerable carbon can be produced, not 
the high carbon steels, and the 

subsequent refining of the iron is 

5. The electrodes must be 
burdened only up to a certain 
amount per sq. cm. of section, so 
that with coke alone the voltage 
must be lowered, and with it the 
furnace efficiency. 

. The first important experi- 
ments with an electric shaft 
furnace were carried out by 


These very extensive experi- 
ments were carried out at the 
request of the Canadian Govern- 
ment in 1906 at Sault St. Marie, 
Ontario, in a furnace built by 
Heroult. As the accompanying 
illustration shows, Fig. 123, the FIG. 123 


furnace differs very much from the Heroult steel furnace, and 
approaches the Girod in principle. The lower part is formed of 
carbon material stamped into place and constitutes one 
electrode, while the other, 1.8 m. (72 inches) long, reaches into 
the shaft from above and can be raised and lowered. The 
shaft is 1115 mm. (3 / -io // ) high. It is slightly conical and is 
built of fire-brick. The current was delivered to the furnace 
at 50 volts pressure. 

Below are given details of these tests which are of the greatest 
interest because the electrode and furnace lining stood up for 
at least several days. 

Test No. 13. The raw materials had the following compo- 

Wilbur Magnetite, 

SiO 2 



Fe 2 3 


55.42 ) 



23 . 04 ) 

A1 2 3 












P 2 5 












Moisture = 14.00% 
Volatile matter = 27.56% 

Fixed carbon = 55 . 90% 

Ash = 2.54% 

S = 0.058% 


S1O 2 =81.71% 

Fe 2 3 = 0.09% 

A1 2 O 3 =14-27% 

CaO = i. 60% 

MgO = 1.11% 

Alkali = 1.22% 


The test lasted 61 hrs., 25 mins. The results were: 

9573- 2 3 kgs. ore smelted 
2973.75 " charcoal smelted 

540.23 " sand smelted 
5832 . " pig iron produced 

462.67 " charcoal used per metric ton 
1726 kw. hrs. used per metric ton 

The analyses of the pig iron and slag were: 
Pig iron: 


=0.04 to 3. 7% 


= 0.012 " 0.075% 


= 0.017 " 0.029% 


= 0.20 " 0.27% 

Gr. C 

= 3-53 " 3-70% 

Total C 

= 3.92 " 5-i8% 


SiO 2 

= 39-30% 

P 2 5 

= traces 


= 27.06% 


= 1.21% 

A1 2 8 

= 18.87% 


= 15-55% 


= 0.35% 


= 0.32% 

In regard to the charcoal used the theoretical amount neces- 
sary was determined as follows: 
In i metric ton of ore there are: 

Fe as Fe 2 O 3 = 387-94 kgs. 

Fe " FeO = 179.21 " 

Slag-forming constituents = 176.00 " 

From this we may calculate the theoretical carbon required: 

387.94 kgs. Fe reduced from FeaOs by C forming 

CO uses = 124.52 kg. C 

1 79 . 2 1 kgs. FeO reduced by C forming CO use . . = 38 . 35 kg. C 

162.87 kg. C 
567. 12 kgs. Feas pig iron with 47% require 26.66 kg. C 

189. 53 kg. C 
This equals - '^ = 334.2 kg. per metric ton. 


The actual amount of charcoal used was 462.67 kg. con- 
taining 462.67 X 0.559 = 258.64 kg. carbon. . It follows from 
this that 334.2 258.6 = 75.6 kg. of the carbon necessary were 
either replaced by the volatile constituents of the charcoal, or 
else the CO produced reduced some of the ore in the shaft of 
the furnace. It is therefore clear that the thermal and chemical 
processes taking place in the shaft are of the same nature as 
those in the ordinary blast furnace, whereby the electric furnace 
is helped. 

The small power consumption is very remarkable for it only 
slightly exceeds the theoretical, when the melting of the slag is 
taken into consideration. From this it is certain that the heat 
is mostly used in the interior of the furnace and that because 
of the heat stagnation near the arc the brickwork will be strongly 
attacked. The test unfortunately had to be discontinued be- 
cause of the electrode not working properly. 

Test No. 14. (Time of test: 64 hrs., 30 mins.) 

The results were: 

4943 . 2 kg. Blairton ore smelted 
2 936.95 kg. charcoal 

338 . 23 kg. limestone 

88.71 kg. sand 

5386.71 kg. pig iron produced 
1968 kw. hrs. used per metric ton 

545 kg. charcoal 


(a) The ore: 

SiO 2 = 6.60% 

Fe 2 3 = 60. 74 I F g 

FeO =17.18)* 

A1 2 O 3 = 1.48% 

CaO = 2.84% 

MgO = 5-50% 

Mn = 0.13% 

P 2 O 5 = 0.037% P =0.016% 

S = 0.57% 

C0 2 = 4-923% 

and loss on ignition. 


(b) Limestone: 

SiO 2 

= 1-71% 

Fe 2 O 3 Al 2 Oi 

= 0.81% 

CaCO 3 

-92. 85% C0 = 5i.96% 

MgC0 3 

= 4.40% MgO =2. 09% 


= 0.004% 


= 0.052% 

{c) Pig iron 



= 3.05 to 5.15% 


= 0.027 " 0.332 


= 0.024 " 0.037 


= 0.07 " o.n 

Graph, car. 

= 2.72 " 3.46 

Total car. 

= 3.54 " 4.16 

(d) Slag: 

SiO 2 

= 33 to 37% 

A1 2 3 

= 9 " 18% 


= 18 " 30% 


= 21 " 30% 


= o.oi " 0.05% 


= 0.4 " 0.9% 


_ - *0/ 

- * 6/0 

Test No. 16. (Time of test: 38 hrs., 20 min.) 
The results were: 

2175.6 kgs. Calabogie ore smelted 

1611.7 " charcoal 
587.9 " limestone 

34.1 " quartz 
3246.0 " pig iron produced 
497.0 " charcoal used per metric ton 
1970 kw. hrs. per metric ton 

(a) Ore: 

SiO 2 = 6.06% 

Fe 2 O 3 =58.00 | _ 

FeO =24.78) 

A1 2 O 3 = 1.00% 

CaO = 0.40% 

MgO = 6.00% 

P 2 O 5 = 0.046% P = o. 02% 

S = 0.17% 

COa = 3-544% 

and loss on ignition. 


(b) Charcoal: 

Moisture = 2.20% 

Volatile matter =20.60% 
Fixed carbon =74.40% 
Ash = 2.80% 

(c) Lime. The same limestone was used as in Test No. 14^ 
No analysis was made of the quartz. 

(d) Pig iron produced: 

Si = 1.22 to 2.03% 

S = 0.006 " 0.008% 

P = 0.047 " 0.093% 

Mn = 0.07 " 0.12% 

Graphitic C =3.87 "4.55% 
Total C = 4.40 " 5.06% 

(e) Slag produced: 

SiO 2 =30.88% 

A1 2 8 = 9-67% 

P 2 O 5 = 0.014% 

CaO =36.14% 

MgO =20.82% 

MnO = 0.14% 

FeO = 0.73% 

S = 1.23% 


The three tests given above show the following consumption 
of power for the production of one metric ton of pig iron: 

Kw. Hrs. 

Charcoal kg. 







Average 1,888 501 

This power consumption is good, exactly as in all the former 
tests, because the charge forms a good heat-insulator. Still, 
this concentration of heat has proved a disadvantage, for with 
the great drop in temperature between the arc and the walls of the 
furnace the limited amount of charge surrounding the arc is not 
enough to absorb it, and the result is a rapid destruction of the 
lining and uneconomical operation. The ascending reduction gases 


cannot lead away the excess heat near the arc through the charge 
to the throat, so that the lower part of the furnace is necessarily 
quickly destroyed by the "stagnant heat." 

The intended prereduction of the ore is brought about, 
although only to a moderate degree, so that the carbon con- 
sumption is still high. The long electrode hanging in the furnace 
is shown to be a mistake because it is continuously exposed to 
mechanical wear, and is also chemically attacked by the sur- 
rounding ore mixture. Because of this delays in operation may 
be caused. 

From all this it follows that the problem of electric ore 
smelting is not to be solved by the Heroult type of furnace, 
because electrode consumption, delays in operation, and the 
lining costs exclude economy. The quality of the metal pro- 
duced, on the other hand, is good. Phosphorus and manganese 
are completely reduced, the slag can be kept low in iron, and the 
production of low sulphur pig iron of any desired silicon is 
possible. As a reducing agent lump charcoal and also peat coke 
can be used. 


Gronwall, Lindblad & Stalhane knew that the amount of 
reducing gases developed was not sufficient to carry the excess 
of heat present near the arcs from the lower part of the furnace 
to the shaft where it could be used economically for preheating 
the charge. They therefore increase the amount of gas by 
forcing into the lower part of the furnace, by means of a fan, 
part of the waste gases drawn from the throat. The amount 
is regulated so that the excessive heat, which -would soon lead to 
the destruction of the lower part of the furnace, is driven into the 

Because of the continuous operating troubles experienced 
with the long Heroult electrode, Gronwall, Lindblad & Stalhane 
used three electrodes introduced at the sides of the lower furnace 
in their early tests. 

A general view of the furnace is shown in the accompanying 


illustration, Fig. 124. In principle it is similar to an ordinary 
small blast furnace, the electrodes taking the place of the 
tuyeres. The extensive smelting tests carried out with all kinds 
of ores allow a definite opinion to be formed as to the practical 
efficiency of this type of furnace, and show that there is no 
more difficulty in making iron with 4 per cent, carbon in the 
electric furnace than in the ordinary blast furnace.* Theor- 
etically, the direct production of the harder steels is also pos- 
sible, but experience has shown that such steels do not have the 
required temperature, and must be tapped while thickly fluid, 
which leads to troubles in operation. 

In regard to construction the furnace has shown that the 
expected advantages are obtained. First with reference to 
avoiding the stagnant heat in the lower part of the furnace which 
would lead to rapid wearing away. In the first test furnaces, 
which were built either with none or a very small shaft, the 
reducing gases escaped at 70 C., but with the new construction 
the gases at the throat have a temperature of 200 C. to 250 C. 
(see Fig. 129), and the radiation loss of the shaft is also equalized 
by these hot gases. 

From this it follows that the lower part of the furnace will 
stand up better during operation, but the efficiency of the 
furnace will not be so great, that is the power consumption 
necessary per ton of pig iron will be higher. 

Second, in regard to preheating and preliminary reduction 
of the ore, while smelting the ore in the electric furnace, having 
no shaft, only pure carbon-monoxide is produced, the waste 
gases in the electric shaft furnace give the following analysis: 

In 1909 In ign 

a Charge of Red Hematite Fe 2 O 3 Charge of Hematite 

CO 2 =45% CO 2 CO H CH 4 N O 

CO =40 27.2 57.5 14.8 o.o 0.5 o.o 

H 2 =15 

b Charge of Magnetite, Fe 3 O 4 Charge of Magnetite, Mch. 16 6* jo 
CO 2 =30% CO 2 CO H CH 4 N 

12.6 71.9 13.0 1.7 0.8 
19.2 59.7 17.6 2.5 i.o 

* See American Electro-Chemical Society, p. 400, 1911. Robertson. 



According to the researches of Bauer and Glaessner, the 
reduction of iron ore by carbon-monoxide begins at about 
650 C., and is most active at about 700 C. 

On the other hand, according to the tests made at Trolhattan 
in 1911, and given by Robertson, the reduction of magnetite 

FIG. 124. 

by carbon-monoxide takes place at as low a temperature as 
300 C. As the above table shows, this furnace gas contains 
about 72% of that gas, so that reduction of the charge by the 
gas rich in CO probably takes place throughout the whole of 
the lower half of the shaft, since the temperatures from the 
official report of the Jernkontoret on the working of the Trol- 
hattan furnace for the month of January, 1911, gives a tempera- 


ture at the foot of the shaft (Point No. i, in Fig. 125) at 534 C. f 
and at point No. 4 at 351 C. The readings at point No. 5 were 
discontinued, but from the other figures given it would appear 
that the temperature at this point is not below 300 C. 

As the gases in the electric shaft furnace leave the throat at 
200 to 250 C., the chosen height of shaft of 5 metres (16' 4.8") 

FIG. 125. The Gronwall, Lindblad and Stalhane furnace, design of 1911. 

is more than sufficient. With the use of the Lange bell alone 
this height can be lowered, and with the use of the Parry cone at 
the same time, the effective height can be considerably decreased. 
The high hydrogen content of the gas comes partly from the 
hydrogen contained in the reducing agents, but partly from the 
moisture in the charge, which is decomposed. Hydrogen does 
not have a special reducing action in the presence of carbon- 
monoxide, which explains the given high hydrogen content. 


This is shown by recent experiments in the production of pure 
hydrogen in large amounts which consist of strongly heating 
iron ore in a muffle furnace, and treating it with water-gas. The 
ore is reduced, yet almost the whole of the hydrogen passes from 
the furnace unoxidized, and is used for heating the furnace. The 
reduced iron is then employed to produce pure hydrogen, by 
passing steam over it. 

That an active prereduction takes place in the electric 
blast furnace is proved by the gas analyses, and the saving 
brought about in this way should be considered in calculating 
the amount of the reducing agents to be charged. This saving 
is based on the ratio of CO 2 to CO in the waste gases, which, 
for example, in the case of magnetite may be 40 CO 2 : 60 CO. 
The gas contains 100 carbon to 140 oxygen, the latter coming 

from the magnetite Fe 3 O 4 , the amount being =35. This 


is to say that the reduction process is based on the formula 
35 Fe 3 O 4 + ioo C. According to this 35 X 3 X 56 parts of 
iron and (ioo X 12) +3 parts of carbon should be charged for 
the production of a pig iron with 3% carbon. If the amount 
of CO 2 in the gases falls below 30%, then there is an excess of 
raw material over the carbon present for reduction, because 
more ore enters the lower part of the furnace, and some additional 
material rich in carbon must be charged. On the other hand, 
if the charge contains too much carbon, then the lowerp art of 
the furnace becomes filled up with it, and some additional lower 
carbon material must be charged. 

In regard to the slag, a singulo-silicate is the best, with the 
formula Si0 2 2 CaO, and the proper amount of lime to produce 
this must be added to the charge. As with 'the ordinary blast 
furnace so also here it is not profitable to run too basic a slag, as 
the power consumption increases more than it should.* In 
this case the slag also very often contains calcium carbide formed 
by the influence of the arc. The power consumption per ton 

* The analyses of slag, according to Leffler, which follow show that these 
have generally been kept more silicious than desirable for the basic lining 
of the hearth. This, however, has been done for the purpose of obtaining 


of pig iron naturally depends on the amount of slag that has to 
be melted, because it must be tapped in a fluid condition. 

Technical knowledge in 1909 was such that only high per- 
centage ores with 65 to 68% iron could be successfully smelted, 
which were as low as possible in sulphur, and which in no case 
gave more than 10% of slag. 

Further progress was made in the 1910 and 1911 tests such 
that ores running as low as 53.25% iron and containing .055 sul- 
phur (Nordmarken coarse washed ore) were successfully smelted. 

Ores with high sulphur should therefore be roasted before 
smelting in order to reduce the amount of lime necessary to be 
added to the charge. This roasting is comparatively easy with 
ores with an acid gangue, an average result with Swedish magne- 
tite showing: 

Before roasting 0.7 % sulphur 

After 24 hours 

o.i % 

With a large electric shaft furnace plant the waste gases can 
be used for heating the roasting furnaces. Certain magnetites 
swell during roasting, do not break up, but change into red 
hematite. This change probably only makes somewhat lighter 
the consumption of reducing material and electric energy, for 
in the ordinary blast furnace 100 parts of magnetite need 100 
parts of coke, while the same amount of red hematite takes 90 
parts of coke. Fortunately, definite figures on this point have 
been obtained for electric furnace work, and are as follows: 

These are taken from the 1910 and 1911 Trolhattan tests. 
In those singled out for comparison ores of about the same iron 
content (65%) were chosen. The first test lasted 2096 con- 
results as closely comparable as possible with the treatment of the same ores 
by the ordinary blast furnace process. 



Si0 2 

A1 2 O 3 TiO 2 FeO 





P 2 6 






















secutive hours and used 1,760,884 kg. (about 1,760 tons) of 
natural magnetite ore. The charcoal used per ton of iron equalled 
415.7 kg. (914 lb.), containing 70.5% C. The second test lasted 
193 hours and used 223,626 kg. (about 223 tons) of magnetite 
ore of which about 87% was roasted. The charcoal used per 
ton of iron here equalled 376.3 kg. (829 lb.), containing 73.5% C. 
The slightly higher carbon in the charcoal content of the latter 
case is perhaps offset by only 87% of the ore, in this case having 
been roasted, thus making the comparison with raw and all 
roasted ore better, and about as it would be if in the one case 
all the ore had been roasted and the charcoal in each case con- 
tained the same carbon content. The reduced amount of char- 
coal used for the roasted ore is about the same as with ordinary 
blast furnace practise, viz. 10%. 

Ores with a basic gangue give great trouble in roasting, for 
the sulphur forms gypsum, and the intended reduction in sulphur 
is prevented, therefore such ores high in sulphur should not be 
used in the electric shaft furnace. 

Fairly rigid requirements are also necessary in the physical 
properties of the ore to be smelted. The most suitable size is 
that of a large walnut, and only a little pure ore should be 
present. Lump ores have, therefore, to be crushed and none 
can be used which give a considerable percentage of fines on 
crushing. This is sometimes a great disadvantage because of 
the brittle character of many magnetites, etc. The reducing 
agent also ought to be about the size of one's fist, as much as 
possible, and fine material can only be used with difficulty and 
in small amount. 

Formerly, i.e., in the earlier tests only charcoal could be 
used, or a mixture of coke and charcoal. Since then, however, 
a 3000 to 3500 HP furnace of the Gronwall, Lindblad & 
Stalhane type has been completed and is in operation at Har- 
danger, Norway, where English Durham coke, carrying about .6% 
sulphur, is being used. This is according to Richards, A. E.G., 
Society, 1911, p. 417, and from private advices from D. A. Lyon. 

In regard to the practical operation in 1909, small charges 
had to be used corresponding to the small size of the furnace, 


that is to say, charges containing about 100 kg. (220 Ibs.) of 
ore. In the larger furnace of 1911 the average charge over a 
6 months' period was 425 kg. of ore (937 Ibs). One-half of the 
ore should be thrown around the outside, and the rest with the 
reducing material and lime in the centre. If the charge were 
placed only in the centre, the 1909 furnace would easily hang, 
ordinarily due to the separation of carbon. Eighty charges 
containing 8 metric tons of ore were smelted in 24 hours, which 
with a 65% ore gives a total output of 5.3 to 5.4 metric tons of 
metal, obtained at intervals of 6 hours. Because of the small 
amount of slag, it was allowed to remain in the furnace, and was 
tapped together with the metal. 

The results confirm those already obtained with the Heroult 
furnace, namely, that the smelting process is the same as that 
of the ordinary blast furnace, so that from a corresponding ore 
any desired pig iron can be obtained by running a suitable slag, 
and regulating the furnace temperature. With a high tem- 
perature the iron contains more carbon, and if at the same time 
a basic slag is run the manganese of the ore is completely reduced, 
and a low sulphur iron is obtained because of the complete re- 
moval of the sulphur in the slag. The silicon content when 
running a basic slag and high temperature decreases, and under 
these conditions a part of the phosphorus can remain unreduced 
in the slag. On the other hand if the slag is acid the manganese 
is partly slagged off, and with high temperatures a high silicon 
iron is obtained. Just the same conditions obtain here, there- 
fore, as in the ordinary blast furnace. In operation it is always 
desirable to produce an iron as low in carbon as possible, which 
is the most favorable for foundry purposes, and also for subse- 
quent refining into steel. 

In regard to power consumption, in the tests ending in 1909, 
280,307 kilograms of iron were produced in 1903.5 hours, during 
5.9% of which no work was done due to troubles with the ma- 
chinery. For each metric ton, 3181 kw. hrs. were used with an 
electrode consumption of 30 kg. or 66 Ib. = (0.015%), and an 
electrode loss of 8 kg. or 176 Ib. = (0.004%). The production 
from the ore was 63.5% and from the charge 60.02%. 



The reducing agent weighed 354. kg. (779. lb.), and con- 
sisted of 41.7% coke and 58.3% charcoal, and a total of 35.41% 
was necessary, which corresponds to a consumption of 28% pure 
carbon. From this data the efficiency of the electric shaft 

FIG. 126. The Gronwall, Lindblad & Stalhane furnace. Latest design 
of 1912. Note lower position of electrode clamp. 

furnace can be calculated. The pig iron may be taken as con- 
taining i% silicon and 3% carbon, which leaves 96% iron, the 
ore as magnetite, and the waste gases as containing 30% C0 2 
and 70% CO. 


ioo kg. of the waste gases contain 30 + 70 = ioo kg. carbon, 
and (2 X 30) + 70 = 130 kg. oxygen. In the reduction of 
magnetite 130/4 kg. Fe 3 4 must be present to supply the oxygen, 


FIG. 127. The Gronwall, Lindblad & Stalhane furnace. Gas circulation 

of 1911. 

and 130/2 = 65 parts of silica to supply the silica. Reduction 
takes place according to the following formulae: 

1. kg. Fe 3 4 + ioo kg. C = 30 COo + 70 CO 


or 13 kg. Fe 3 4 + 40 kg. C =12 C0 2 + 28 CO 

2. 65 kg. SiO 2 + ico kg. C = 30% C0 2 + 70% C. 

This gives (12 X 12) -f (28 X 12) = 480 kg. carbon, which ac- 
cording to the analysis of the gas gives (12 X 44) + (28 X 2 8) = 

L . 1312 X ioo 

1312 kg. gas, that is to say, i kg. carbon gives - 




kg. gas. (i Ib. carbon gives 2.73 Ib. gas.) 


13 kg. of magnetite require 40 kg. carbon for reduction, so 
that for the smelting of 13 X3X 56 = 21 84 kg. iron 40 X 12 = 

u r /: i Q^o X 480 

480 kg. carbon are necessary, or for 960 kg. iron - - = 

, 2184 

2 10.99 kg. 

65 X 28 X 364 kg. silicon reduced from silica require 100 X 
12 = 240 kg. carbon or 10 kg. silicon require 6.59 kg. carbon. 

For carburizing the iron, 30 additional kg. of carbon are 
necessary, so that the total requirement of carbon necessary for 
the production of i metric ton of pig iron amounts to 210.99 + 
6.59 + 30.00 = 247.58 kg. (545.8 lb.). 

From this there is formed (210.99 + 6.59) X 41.15 = 
594.72 kg. waste gases (1311.1 lb.). 

With an output from the charge of 60%, 960 kg. iron require 

, 960 X ioo 960 X ioo 

a charge of - = 1600 kg., with = 1325.71 

oo 72 X 4^- 

kg. Fe 3 O 4 . This will produce 1600 1325.71 = 274.29 kg. 
slag from which ~ = 21.43 kg. silica are reduced and 

enter the iron, leaving 274.29 21.43 = 252.86 kg. (557.4 lb.). 
Heat requirements. The combustion of i kg. carbon pro- 
ducing the waste gas analysis given above creates (0.3 X 8080) -f- 
(0.7 X 2470) = 4153 cals. 

Reduction of 960 kg. iron from Fe 3 O 4 = 960 X 

1648= 1,582,080 cals. 

Reduction of 10 kg. Si from SiOa 10X7829= 78,290 

Smelting and overheating of 1000 kgs. pig 

iron = 1000X280= 280,000 

Smelting and overheating of 252.86 kgs. slag 

252.86X595= I5045 2 " 

Heating of 594.72 kgs. CO 2 and CO to 200 C. 

594.72X200X0.245 = 29,145 " 

2,119,967 cals. 

Heat Supplied: 

Combustion of 217.53 kgs. carbon =217. 53 X 

4153 = 903,610 cals. 

Leaving to be supplied by the electric current. . . . 1,216,357 

Total 2,119,967 cals. 


The theoretical amount of power necessary for 1000 kg. of pig 

iron is 


As 3181 kw. hrs. are required daily per metric ton, the 

efficiency is - - = 44%. This low efficiency obtained 


in the 1909 tests is for a small test furnace run with constant 
supervision. The remaining 56% is lost by radiation and cooling. 
In this respect measurements have shown that the water cooling 
of the three electrodes carried away about 120 kw. hrs. per hour. 
This gives a total loss of 1903.5 X 120 for the entire smelting 

test, which equals 228,420 kw. hrs. or ~ = 815.7 kw. hrs. 

/ / O 

per metric ton, which corresponds to ~ - = 25.6% of 

the electric energy supplied. Through radiation alone 56 
25.6 = 30.4% of the energy is lost. 

It should now be considered 
whether and by how much the 
efficiency can be increased with 
a larger plant. Water cooling 
will still have to be used, and in 
this respect the efficiency can 
scarcely be increased. Apart 
from this the high water con- 
sumption, amounting to about % 
gallon per second (2 liters) is a 
disagreeable addition. On the 
other hand the radiation loss 
would be smaller because the 

cubic contents increase faster than the radiating surface of the 
furnace. Most important, however, is the fact that the smelt- 
ing time per ton of iron will be lowered, and therefore the radia- 
tion per ton of metal will be considerably smaller with the 
increase in smelting efficiency. 

Graphite electrodes will increase the smelting efficiency for 
they are better conductors than those of carbon, and although 

FIG. 128. Modified gas circu- 
lation of 1912. Gronwall, Lind- 
blad & Stalhane furnace. 



they have a higher thermal loss (Chapter VI, Part I), yet this is 
more than equalized by the increased efficiency. At the furnace at 
Falun carbon electrodes were used, for there is no plant in Sweden 
making graphite electrodes. This 
dependence on electrode plants is 
necessarily very disadvantageous for 
all countries not having them. Ex- 
periments should be made to in- 
crease the life of the electrodes by 
mechanical means as much as pos- 
sible, or the electrode consumption 
is proportionally high. It will not 
be much lower with a large furnace, 
as the electrodes are attacked be- 
cause of their contact with the 
ore. Finally the consumption of 
reducing material is very much 
higher than it should be theoreti- 
cally, which of course is also not 

Below are given some details of 
recent test runs in larger furnaces, 
and it may be seen how these theo- 
retical considerations have worked 
out in practise. 

With regard to the run from 
Nov. 15, 1910, to April 9, 1911, in 
the newer Gronwall, Lindblad & 
Stalhane or Swedish Ludvika Elek- 
trometal type furnace, 1882.496 kg. 
(about 1882 tons) of iron were 
produced in 3501.9 hours, during 
about 4.4% or 153.7 hours of which 
no work was done due to troubles 
with the apparatus. For each metric FlG - I2 9- -Temperatures in 

, . . , the Gronwall. Lindblad & 

ton, 2391 kw. hrs. were used with Stalhane electric pig . iron fur . 
an electrode consumption of 10.28 nace. 


kg. (22.6 Ibs.) gross, and 5.27 kg. (n.6 Ibs.) net, per ton of 
iron produced. The per cent, of iron in the ore was 61.54% 
and the per cent, iron in the charge 57.00%. The reducing 
agent weighed 418 kg. (920 Ibs.) per ton of iron produced and 
consisted of charcoal only, having a carbon content of 80.14%. 
The pig iron may again be taken as containing i% silicon and 
3% carbon, which leaves 96% iron. The average of the gases 
produced consisted of 23% CO 2 , 60% CO, 10% H, 2% CH 4 , and 
5% N. From this data the efficiency of the electric pig-iron 
furnace may again be calculated as before, and in this case the 
efficiency is considerably higher, being about 59%. 

After the furnace at Trollhattan was shut down from June to 
September, 1911, in order to make such changes as the operation 
of the furnace had demonstrated would be beneficial and such 
repairs as were necessary, the furnace was again put into com- 
mission. During the run from Sept. 3 to Sept. 30, 537.9 tons 
of pig iron were produced. For each metric ton of pig iron 1 749 
kw. hrs. were used with an electrode consumption of less than 
10 kg. (22 Ibs.) gross and 5 kg. (n Ibs.) net. The iron in the 
ore was 67.65% and the iron in the ore and lime was 65.02. The 
reducing agent weighed only 339.9 kg. now (748 Ibs.), consisting 
of charcoal. With the same carbon content as before, 72%, this 
equals 245 kg. or 24.5% pure carbon. From this data the 
efficiency of the furnace can again be calculated and figured out 
to 80.5%. This corresponds to an output of over 5 tons of pig 
iron per kilowatt a year. The above efficiency corresponds 
favorably with the Swedish charcoal blast furnace of 82% and 
with 70% the usual coke blast furnace. 

As a conclusion it may be said that the Gr on wall, Lindblad 
& Stalhane electric shaft furnace is probably the first electric 
furnace in which ore has been smelted in some degree commer- 
cially. The weak point has been the furnace roof, which, in the 
1909 furnace, either showed such small durability and therefore 
made continuous operation impossible, or else had to be cooled 
so strongly that the efficiency of the furnace suffered consider- 
ably. Further the close limits allowable in the chemical and 
physical composition of the charge, and the large electrode 


consumption show that the furnace can be employed only under 
especially favorable operating conditions. The principle is first- 
rate, especially with regard to making the roof of the lower part 
of the furnace more durable, as far as possible without the use of 
water cooling, and so increasing the furnace efficiency. In this 
place the Lyon experiments conducted at Heroult, California, 
with the Noble furnace, should be mentioned. The following 
details are taken from a paper by Otto Frick, in Metallurgical 
and Chemical Engineering, December, 1911, on "The Electric 
Reduction of Iron Ores." 

The Noble furnace is of the same type as that at Trollhattan, 
and like it in all essential points. This, however, is not the 
result of mutual understanding or communication. An illustra- 
tion is given of the furnace at Heroult by Fig. 130. This 
furnace has now (1912) been rebuilt seven times. It has three 
single-phase transformers, each of 750 kilo volt amperes, con- 
nected to a three-phase system of 2200 volts and 60 cycles. 

The low tension current is supplied to six graphite electrodes. 

These electrodes enter into the charge as far as possible, and 
in this respect the practise differs from that at Trollhattan, where 
a space is left between the electrodes and the charge. The 
pressure of the charge on the electrodes is very nearly equal to 
their breaking strength, so that the additional force arising from 
a sudden descent of the charge easily causes their breaking at 
the conical screw joint. This strain can be reduced approxi- 
mately 30% by lowering the inclination from 35 to 20, and 
much improvement can be made in the joint. 

No accurate figures are at hand as to the power consumption, 
but it has been stated by the manager of the plant that it has 
averaged 1940 kw. hrs. per ton. 

With regard to gas circulation it has been found unnecessary 
to use any in the Noble furnace, where the electrodes penetrate 
the charge far enough to prevent arcing, so long as they remain 

The question of the smelting of ore in the electric shaft 
furnace can only be considered solved when the following require- 
ments are met: 


1. Water cooling lowered as much as possible. 

2. The electrode consumption lowered, and the electrodes 
done away with as much as possible. 

3 . The radiation loss lowered by the smelting efficiency being 
raised as much as possible. 

4. The waste gases composed of only pure carbon-dioxide, 

FIG. 130. The Lyon furnace in California. 

of suitable temperature, in order to lower the amounts of electri- 
city and reducing material necessary. 

Requirements i and 2 are very closely connected. So long 
as electrodes are used, cooling of the electrode heads cannot be 


avoided, which brings about great heat loss, and necessitates a. 
complicated furnace construction. Further great durability of 
the furnace lining is only possible if the high initial temperature 
of the arc is avoided, and the most suitable moderate tempera- 
ture used. This requirement is only met by the induction furnace,, 
for on the one hand electrodes are not used at all; and, on the 
other hand, as the experiments with the hearth induction furnace 
have shown the furnace lining is hardly attacked at all when 
smelting ore. An induction furnace with a wide hearth and a 
shaft built above is the one to claim the greatest interest for 
smelting ore, and so much the more that it can be operated at 
the highest temperatures if required. Any height of shaft can 
be chosen, so that the waste gases can be efficiently used for 
prereduction and preheating of the charge. 

The radiation loss decreases with a larger furnace for the 
induction as for other furnaces. 

In regard to the requirement that only pure carbon-dioxide, 
at a suitable temperature, should be given off, as waste gas, it 
is well known that carbon-monoxide loses the ability to reduce 
ore when a certain percentage of carbon-dioxide has been formed. 
It is therefore theoretically impossible to have a product of pure 
carbon-dioxide when charging ore and fuel. 

The complete utilization of the waste gases is therefore only 
possible if they are burned afterwards, and used as much as. 
possible for preheating the ore. 

This preheating favors smelting only in that reduction is 
made more easy by an increase in the degree of oxidation, and 
also because the sulphur contents are lowered so that a low sul- 
phur iron can be obtained without the addition of more flux to the 
charge. With finely divided ores the roasting also brings about 
a certain amount of agglomeration so that under these con- 
ditions fine ores, concentrates, etc., can be smelted in the electric 
shaft furnace. The heating and agglomeration of fine ores, if 
sufficient fluxing material is present, can be carried out in a 
revolving cylindrical furnace, the ore being charged wet just as it 
comes from magnetic separation for instance. Such agglomerat- 
ing plants are already in satisfactory operation. 


The use of coal dust firing which is recommended for the 
heating of these furnaces is unsuitable, as it gives rise to high 
fuel costs, and the ash of the coal makes the ore higher in slag- 
producing material so that it is more unsuitable for electric fur- 
nace work. It is, therefore, necessary to operate with waste 
gases, and an addition of producer gas should only be made when 
the difference in price between fine and lump ore is sufficiently 
great to bear the increased cost in fuel needed for the production 
of the producer gas. 

Such an ore, however, greatly preheated, cannot be charged 
directly with the reducing material, as it is immediately reduced, 
forming carbon-monoxide, and so reduces the furnace efficiency. 

The greatly heated ore must be charged alone, and the 
reducing material introduced in the hearth of the furnace at the 
deepest zone of the shaft. The physical condition of this re- 
ducing material is not important, if solid, the most suitable size 
is fine grained. Very small fuels and even valueless waste can 
be used with complete success. 

Also fluid- reducing materials such as tar, petroleum, and 
oil residues of all kinds can be used. This is of special interest 
to those countries which at present must import coke or charcoal, 
because these liquid fuels due to their high heating value and 
low ash contents are brought in at much more favorable freight 
rates. Finally gaseous reducing agents of all kinds can be used, 
such as producer gas. The carbon-dioxide should be as low as 
possible, and if fuels with much moisture are used, such as turf, 
brown coal, etc., the gas should be cooled as thoroughly as possible 
to remove the moisture The troublesome precipitation of tar 
experienced in the cooling of producer gas is no disadvantage to 
the electric furnace, as opposed to other furnaces, for the tar can 
be collected, dried in centrifugal machines and used in the furnace 
as a reducing agent. 

The carbon-monoxide or the solid liquid or gaseous materials 
used easily reduce the highly heated charge, and give warm waste 
gases rich in carbon-monoxide, which can serve to preheat more 
ore charges if air is added to combine with the combustible 
constituents . 


In this way it is possible to considerably reduce the consump- 
tion of reducing material, and to come very near the theoretical 
minimum; which, in the case of magnetite and the production 
of a pig iron with 3% carbon, is 143 + 30 = 173 kg. of carbon 
(381.4 Ibs.) per metric ton. The best figures reached so far as 
already mentioned are 245 kg. of pure carbon when making a 
pig iron with 3.64% C. 

After nearly a year of further experience (215 days) in operat- 
ing the furnace at Trollhattan, Leffler and Nystrom contributed a 
supplementary report of 98 pages, to the meeting of the Jern- 
kontoret at Stockholm, on May 31, 1912. It is not possible to 
do this report justice here by any abstract of it, still it may be 
instructive to mention some of the improvements recently made. 
The new gas circulation system was altered to better dry the 
gas returned to the furnace. Fig. 128 shows the latest design 
and is but little different from its predecessor shown by Fig. 
127. The cooler acts on the condenser system and requires 100 
liters of water per minute to reduce the temperature of the gas 
so that its moisture content is reduced from 4 grams per cubic 
meter to .5 gram. 

Both high and low grade ores were used in this run, so that 
the furnace output dropped to about 15 tons daily from its 
normal capacity of 20 tons. This run again demonstrated that 
economical operations need a rich ore. 

Fig. 129, which is reproduced from the July, 1912, Metallurgi- 
cal and Chemical Engineering, shows the temperature and 
reaction in the furnace shaft. This abstract goes on to say: 

The temperatures of iron and slag issuing from the furnace 
varied as follows: 

Iron 1230 to 1420 C. 

Slag 1290 to 1460 C. 

A large table gives the temperatures taken at 8 points in the 
shaft, just inside the wall and in the middle; also the percent- 
ages of CO 2 in the gases at thee different points. 

There are various not very important irregularities in the 
figures, but the general average shows temperatures up to 985 
in the middle at the lower part of the shaft and 585 half way 


up; while near the wall it is 420 to 565 at the lowest hole and 
down to 15 at the highest. 

The measured percentage of CO 2 shows that reduction takes 
place ordinarily only one-quarter way up the shaft at the sides 
and a little over one-half way up in the centre. 

The extent of the zone of reduction by CO is clearly shown 
in Fig. 129, in which also some temperatures are indicated. 

Cooling Water. The contacts and jackets through which the 
electrodes worked were water cooled. The heat carried away 
thus varied from 172 to 288 kw., or 10.47 to 19.30 (average 
14.50) per cent, of the power used. 

Thermal Balance. The heat balance per 1000 kg. of pig iron 
is worked out for the four weeks, Sept. 3 to Oct. i, 1911, in which 
the average power used (high tension side) was 1407 kw., and 
the power consumption 1749 kw. hours per ton of pig iron; the 
ores worked were the rich Tuolluvaara ores. The heat balance 
is, per kg. of iron: 

Combustion C to CO 2 567 calories 

Combustion C to CO 381 

Electric energy 1504 

2452 calories 

Consumed in reductions 1620 calories 

Decomposition of limestone 35 

Evaporation of water 24 

Sensible heat in throat gases 26 

Sensible heat in slag 75 

Sensible heat in pig iron 300 

Cooling water 195 

Lost in transformers 43 

Lost in conductors 44 

Radiation and conduction 90 

2452 calories 

The authors then make some interesting calculations, the 
results of which are, in brief, as follows: The gas kept in cir- 
culation was 2.28 times the gas normally produced and escaping. 
Assuming this gas to enter the furnace at 22 and to enter the 
shaft at 1000, it carried into the shaft as sensible heat 343> IlS 


calories per ton of iron, or 22.9 per cent, of all the heat electrically 
generated in the crucible. Since it carried with it 22.5 kg. of 
water vapor and 174 kg. of C0 2 , both of which are decomposed 
by the glowing carbon, the net heat absorbed in these decom- 
positions is 160,283 calories, or 10.7 per cent, of the electric 
energy used. The gas circulation therefore transferred physically 
and chemically 33.6 per cent. = 1/3 of the electrical energy used 
from the crucible into the shaft of the furnace. 


Pig iron, that is the iron and carbon alloy, produced in the 
electric or ordinary blast furnace or in any way, contains other 
constituents such as silicon, manganese, sulphur, phosphorus, 
copper, arsenic, etc., which come from the charge. Some of 
these elements, such as copper and arsenic, are easily reduced 
from the ore and enter the metal, and cannot be removed econom- 
ically by any metallurgical operation. The other elements, 
such as silicon, manganese, sulphur, and phosphorus, can be 
partly eliminated in the blast furnace and slagged off, and they 
can also be separated more or less from the finished metal by 
later metallurgical operations. 

If the amount of one of these constituents is to be lowered 
in order to make the metal more suitable for any special purpose 
it is spoken of as a refining of the metal. Therefore a lowering 
in the carbon percentage of the metal is also to be con- 
sidered as a refining. The refining process can be of various 
kinds, reducing and oxidizing, or consist of simple reactions 
such as: 

FeS + Mn = MnS + Fe. 

If an electric furnace is to be suitable for refining, then all 
processes, whether oxidizing or reducing, must be practicable; 
above all it must allow the carrying out of all metallurgical 
operations, such as are now used in the open hearth, converter, 
etc. The electric furnace, and this must be specially pointed 


out, should not be different from an ordinary furnace, except 
that the heating is electro-thermal. The electric furnace, as 
such, except for the lining, should have no influence on the 
chemical composition of the bath of metal. Arc furnaces do 
not correspond altogether to these requirements, for an influence 
of the electrodes on the bath cannot be avoided even with 
careful operation. 

Electric heating of the furnace has the great advantage that 
the influence of the hot gases on the charge, which is present 
in the furnaces used now, is excluded, so that work can be 
carried out with an oxidizing, neutral, or reducing atmos- 
phere at will. Even the maintenance of a neutral or reduc- 
ing atmosphere is not only very difficult with the present 
furnaces but really impossible, except with crucible and muffle 

The induction furnace completely meets these requirements, 
for in it a reducing or oxidizing atmosphere can be obtained as 
desired. With the arc furnace on the other hand, reduction 
processes take place very well, but oxidation processes only 
slowly, due to the reducing action of the electrodes, and with 
.an increased use of oxidizing material there is more electrode 
consumption. Otto Thalner gives expression to this in his 
address before the " Oberschlesischen Bergwerksverein deutscher 
Chemiker," 1909, when he says: "The arc furnace is indeed a 
good reduction furnace, but a bad refining furnace." An electric 
furnace, however, to answer all requirements should allow reduc- 
tion and oxidation processes to be carried out equally well, and 
this should be pointed out before anything is said about the 
metallurgy of iron and steel, the influence of impurities, or the 
refining of the metal. 

Phosphorus. This exists in the iron in the form of phosphide 
of iron which dissolves in the metal bath without difficulty up 
to 1.7% Phos., forming mixed crystals. Phosphorus segregates 
in both pig iron and steel, for the phosphide has the comparatively 
low melting point of 910 C. For instance, in gray foundry 
iron the well known separated bean-shaped pieces are sometimes 
found, which give the following analyses: 





Bean-shaped pieces 
Solid piece near the beans . . 

1-30% P 
0.60% P 

1-30% P 

o.55% P 

1.00% P 

0.50% P 

Analytical proof of its segregation in steel is given in the 
next section under Sulphur. 

If a section is cut from a steel high in phosphorus, polished 
and etched with a solution of copper-ammonium-chloride, by 
Professor Heyn's method, the places rich in phosphorus will be 
colored dark, and one is in a position to determine the segregation 
in the material. As segregated material has considerably lower 
physical properties than normal material, a low phosphorus 
should be specified if a high quality is desired, so that if ordinary 
high phosphorus material is to be used for making high quality 
steel, it must be dephosphorized. In order to do this an Ameri- 
can has proposed to destroy the phosphide by the addition of 
another element according to the equation: 

Iron phosphide + metal = iron + metal phosphide. 

Naturally a metal must be chosen that, in the form of phos- 
phide, does not alloy with the iron but goes into the slag. Such 
reactions are theoretically possible, and have been carried out 
practically to a small extent. Even the silica holding desul- 
phurizing slags of the electric furnace show a certain content of 
phosphides, which can be easily recognized by the garlic-like 
smell when the slags are moistened with water, but this method 
of dephosphorizing has not, so far, become of practical import- 

The removal of phosphorus is only possible with certainty, 
at present, when the phosphorus is oxidized to phosphoric acid, 
combined with lime, and removed as slag. The phosphorus is 
oxidized at low temperatures before the carbon, but at higher 
temperatures only after the removal of the carbon from the 
bath. One can therefore dephosphorize high carbon charges 
without having to previously completely remove the carbon. 
In this case the temperature of the bath should be kept low, an 


easily fusible basic slag charged rich in oxide of iron, and at the 
end of the dephosphorization immediately tapped. To a certain 
degree this method of dephosphorizing requires considerable 
care and experience for complete success, although the melting 
of high carbon heats is economical. 

Dephosphorization is more certain with low carbon and very 
high temperature, and at the same time strongly oxidizing and 
basic slags. Further, the oxidation of the phosphorus can be 
brought about as well by the oxygen in the ores as by that of 
the air. The maintaining of a basic slag naturally requires 
that the work be done on a basic hearth. 

Sulphur. Sulphur can exist in steel as MnS, as well as FeS. 
The latter can alloy with liquid iron while the former does not 
alloy with the liquid metal, and is therefore only present in the 
form of included material. If the bath of metal is allowed to 
stand long enough, then the MnS will rise to the surface because 
of its lower specific gravity, and can be drawn off. This is not 
possible with the remaining FeS which remains alloyed with the 
liquid metal. For this reason the slags which separate from 
the metal, for example from basic Bessemer iron, ' in casting 
ladles, or mixers contain a high percentage of sulphur and also 
manganese, present for the most part as MnS, while the amount 
of iron is not so great. This is shown by the following average 
analysis of ladle slag: 

Iron 6% 

Manganese 4 2 % 

Sulphur 10% 

If these slags come lower in sulphur, then oxidation of the sulphide 
of manganese by the air or by included oxide has taken place. A 
high manganese, however, is always a characteristic of these 
slags so that, after a preliminary roasting, they can be used in 
the blast furnace as an ore of manganese. 

Sulphur is harmful to pig iron, steel, and wrought iron, the 
reason probably being the low freezing point of FeS, whereby 
during the cooling of the bath of metal it segregates to the centre, 
and also brings about the red short character of high sulphur 


It is therefore necessary to desulphurize the iron as much as 
possible before it is made into steel, a process that is carried out 
by the addition of ferro-manganese to the liquid bath, if there 
is not enough manganese already present. The sulphide of iron 
is then, decomposed according to the equation: 

FeS + Mn = MnS + Fe 

and if sufficient time is given the MnS rises to the surface of the 
bath into the slag and can be removed. The process only takes 
place smoothly if a considerable excess of manganese is used. 
Even in this case, however, no total desulphurization is possible. 
The sulphur can only be lowered to a certain degree, about 
-5%> which is still considerably too high for special quality 

In liquid pig iron or steel, rich in manganese, that has stood 
long enough before pouring, the sulphur is to be thought of as 
being present exclusively in the form of FeS. If a microscopic 
section is taken from high sulphur material, polished and etched 
as described under " phosphorus," and the dark segregation 
places examined, then a considerably higher sulphur content is 
found than in the ground mass, but only the same manganese. 
If the segregation were a question of the separation of MnS 
then, with an increasing sulphur content, there would also be 
noticed an increase in manganese, which is not the case. 

Below are given some analyses: 

First material: 

Mn% S% P% 

Pure ground mass 0.48 0.067 0.050 

Segregate 0.48 0.182 o.ioo 

Second material: 

1 . Very black segregate . 0.30 0.097 0.155 

2. Gray segregate 0.30 -55 0.079 

3. Pure material o . 30 o . 040 o . 047 

The elongated sulphide inclusions often seen under the 
microscope that are usually assumed to be MnS are perhaps 
nothing more than inclusions of sulphur holding slag, which 
during the rolling of the hot ingot were not yet solidified in its 


interior and therefore were rolled out. Also the general appear- 
ance of red short material during forging and rolling inclines 
one strongly to the opinion that the Fe FeS alloy separates 
between the crystals but not inside the crystals themselves. 

As already mentioned liquid steel produced in the ordinary 
way and therefore fairly high in sulphur must be further de- 
sulphurized for the production of good quality steel. For this 
purpose the electric furnace is suitable. The following two 
processes are those mostly used for desulphurizing in the electric 
furnace, and both take place most energetically at high tempera- 
tures. They also both require the use of a neutral or reducing 
atmosphere in the furnace and the melting of a strongly basic 

In order to make these basic slags easily fusible additions of 
fluor-spar, and quartz in the form of sand, are made. 

(1) The use of the chemical reaction FeS + CaO + C = 
Fe + CaS + CO. 

For carrying out this reaction, therefore, the help of carbon 
is necessary, and the process can be operated very satisfactorily 
in the arc furnace, due to the favorable influence of the electrodes. 
Carbon must be added to the bath, and for this reason the 
process is used only when it is a question of the production of 
high carbon steels. In melting very soft steels, one must either 
take into account a certain carbonization of the bath and later 
remove the carbon, or else be satisfied with a less complete 
desulphurization. Even if the carbon is only thrown on the 
slag covering from time to time, a certain absorption of carbon 
by the bath cannot be avoided. 

(2) The use of the chemical reaction between silicon and 
sulphur whereby SiS is produced which escapes as gas. 

FeS + Si = Fe + SiS. 

If at the same time a lime carrying slag is formed on the 
metal bath a further desulphurization takes place according to 
the equation: 

2 FeS + 2 CaO -f Si = 2 Fe + 2 CaS + SiO 2 . 

The CaS is removed as slag. It is interesting to know that 


both reactions take place almost quantitatively so that scarcely 
more than the theoretical amount of ferro-silicon must be added 
to the bath, and if desired a low silicon steel can be produced. 
The process is often used in the induction furnace and has the 
advantage that it can be used equally well for high and very low 
carbon heats. The reaction gives a very fluid slag because of 
the increase in the amount of silica. 

Moreover, fluor-spar is also an equally good desulphurizing 
agent when ferro-silicon is used, according to the equation : 

2 FeS + 2 CaF 2 + Si = Fe +2 CaS -f SiF 4 . 

Further, in regard to desulphurization by means of silicon 
in the electric furnace a great many theoretical reactions have 
been suggested, a small selection from which is given below. 

(a) With the use of burned lime. 

1. 2 CaO + SiS - CaS -f Si0 2 + Ca -, but there would 

Ca + Fe S = CaS + Fe. 

2. 2 CaO + 2 SiS = 2 CaS + Si0 2 + Si -, but there would 

Si + FeS = SiS + Fe. 

3. 2 CaO + FeS + SiS = 2 CaS + Si0 2 + Fe. 

The SiS in all these equations is thought of as being pro- 
duced by 

Si + FeS = SiS + Fe. 

Also they all represent the same reaction, namely: 
2 CaO + 2 FeS + Si = 2 CaS + Si0 2 + 2 Fe. 

The slag will be made thinly liquid by the silica produced 
and, in this reaction, i sulphur requires X silicon. 

(b) With the use of fluor-spar, 

1. 2 CaF 2 + SiS + FeS = 2 CaS + SiF 4 + Fe. 

The SiS is produced by the equation FeS + Si = Fe+SiS. 

2. 2 CaF 2 -f 2 SiS = 2 CaS + SiF 4 + Si. 
The Si would decompose more FeS. 


Both reactions therefore mean the same, namely: 
2 CaF 2 + 2 FeS + Si = 2 CaS + SiF 4 + 2 Fe. 

The slag will become more basic, that is thicker, and by this 
reaction also i sulphur requires ]/2 silicon. 
Resume of the equations. 

(a) 2 CaO + 2 FeS + Si = 2 CaS + Si0 2 + 2 Fe. 

(b) 2 CaF 2 -f 2 FeS + Si = 2 CaS + SiF 4 + 2 Fe. 

The reactions are the same except that in one the oxide, in 
the other the fluoride, is the reagent; and by both processes the 
same amount of silicon is necessary. 

Desulphurization by the alternate reaction between FeO 
.and FeS and the formation of S0 2 cannot be carried out with 
fluid metal to complete success, and for this reason it is only 
suitable for such cases where complete desulphurizaion is not 

Silicon. The good influence of a certain silicon content in 
gray pig iron and gray iron castings is well known. To a certain 
degree too high silicon in the pig iron is a disadvantage for gray 
iron castings, particularly for the larger ones, as it brings about 
a coarsely crystalline structure, and therefore makes weaker 
castings. On the other hand the silicon in iron or steel can 
easily be raised by the addition of ferro-silicon to the molten 

Silicon can be removed from molten iron and steel by oxida- 
tion, as well by means of ore as by the oxygen of the air, a process 
that naturally takes place more easily on a basic than on an acid 
hearth. The silicon burns before the carbon if the temperature 
is low, at higher temperatures it is only removed completely 
when the carbon is already partly oxidized, while at high tem- 
peratures the silica in the slag is again reduced by the carbon in 
the bath. 

Copper and Arsenic. Neither of these elements can be 
removed economically at present from the bath, so that if a 
certain copper and arsenic content is required in the finished 
material, an appropriate mixture must be charged. 

Carbon. The carbon can be removed from the bath by the 


oxygen of the ore or by that of the air, with the formation of 
carbon-monoxide. If the refining is carried out by means of 
ore, then iron is reduced, a process that requires heat. It is 
probable that thermetal may dissolve a certain amount of carbon- 
monoxide, for iron heated in a stream of nitrogen shows a melting 
point of 1506 C., but when heated in a stream of carbon-mon- 
oxide only 1406 C., a phenomenon that is explained by the 
assumption that carbon-monoxide alloys, at least partially, 
with iron. On the other hand low carbon steel baths easily 
take up carbon whether the latter is added in a solid, liquid, or 
gaseous condition, either in the elementary form or as carbon 
containing alloys such as ferro-manganese, etc. 

Oxygen. Oxygen may occur in steel combined with other 
elements, for example, as CaO, SiO 2 , MnO, Al 2 Os, etc. 
These oxides are only mechanically mixed, not alloyed with the 
steel, and they are usually classed as " inclusions." Such in- 
clusions are undesirable in high quality steels for they loosen the 
structure, and so lower the physical properties. In addition to 
this, however, steel can contain oxygen in the form of ferrous 
oxide, and such a constituent is especially to be feared for it 
alloys with the metal, and, like sulphur, brings about red short- 

It is possible to remove this ferrous oxide from the metal by 
chemical means, reducing it by other elements according to the 

FeO + X = XO + Fe. 

Elements that can serve as reducing agents for ferrous oxide 
must answer the following requirements: 

(1) They should not bring about any development of gases 
in the reduction, for then the metal does not cast quietly, and 
opportunity is given for the formation of gas inclusions. There- 
fore reduction by means of carbon or carbides, electrodes, etc., 
is bad, because the formation of carbon-monoxide is the result. 

(2) They must have a high volatilization temperature. 
Therefore the alkali metals are bad to use, for they escape from 
the bath for the most part as gas without bringing about reduc- 


tion. This is altogether apart from their strong attack on the 

(3) They must easily reduce the ferrous oxide, and for this it 
is necessary that the metal should dissolve in the bath. In this 
way only is a complete contact and reaction possible. 

(4) They must easily slag off, and separate from the metal. 
Manganese, silicon, aluminum, etc., are generally used as 

reducing agents, and, recently, for producing high quality steels, 
certain alloys of silicon with calcium, magnesium, manganese, 
and aluminum. At the same time vanadium and titanium should 
be mentioned, for their influence ought to be, in the first place, 
very strongly reducing on the last traces of ferrous oxide. 

For ordinary purposes f err o -manganese is mostly used for 
deoxidation. Its reaction with ferrous oxide, however, only 
takes place very slowly, and if the deoxidation is to be moderately 
satisfactory a considerable excess of manganese must be added 
to the bath. For this reason only high manganese material can 
be produced which is not applicable as high quality steel for 
different purposes. The slow influence of the ferro-manganese 
is caused by the alloy having to become dissolved before it can 
alloy with the metal. Solid manganese must first melt in the 
bath before it can carry out its deoxidizing action. Because of this 
the deoxidation process would be accelerated if liquid ferro-manga- 
nese were added, and by using this method the amount necessary can 
be considerably reduced, as the loss of manganese in the shorter time 
is smaller. In the electric furnace, where the ferro-manganese 
works in a neutral atmosphere, the minimum amount can natu- 
rally be used for deoxidation, because the alloy has opportunity 
to react on the bath for a long time without danger of being 
burnt by hot gases. Also the necessary excess of manganese in 
the bath can be lowered, as the manganese can work on the 
bath without trouble. A disadvantage of this method of de- 
oxidizing by means of ferro-manganese is that, with the neces- 
sarily large amount of alloy used, the carbon which is unavoid- 
ably present in the blast furnace alloy also takes a part in the re- 
action. It follows that the bath should be allowed to stand for 
a long time after the ferro-manganese addition in order to allow 


the gas to escape. This gas removal is, however, only complete 
if the bath has been given some opportunity to take up silicon. 
Unfortunately, there is no clear explanation for the influence of 
the silicon. It either reduces the carbon-monoxide dissolved 
in the bath, or else it makes the metal able to unite with the 
gases, especially the carbon-monoxide. The latter view is the 
more probable, for it has been mentioned that iron has a very 
low melting point when exposed to heat in an atmosphere of 
carbon-monoxide, which is easily explained by the theory of 
the existence of an iron-carbon-monoxide alloy. 

Silicon is moreover a very strongly deoxidizing material, and 
scarcely more has to be used than the amount theoretically 
necessary. The silica easily goes into the slag, and there is no 
production of gas, as the small amount of ferro-silicon used adds 
practically no carbon. Heats deoxidized by means of silicon 
can be cast quietly and easily for the reasons just mentioned. 

Aluminum is also an effective reducing agent, but there is 
the disadvantage that alumina is produced which, on account 
of its high melting point, does not slag off completely and some 
remains as a fine net-work in the metal, lowering the physical 
properties of the latter. In the production of high quality 
material the use of aluminum is therefore not to be recommended, 
above everything no aluminum should be used while pouring 
into the moulds, for then much less heat is present for melting 
the alumina than in the furnace. 

Recently alloys of vanadium and titanium have been recom- 
mended, the latter produced by the Goldschmidt reaction. 
They are very effective, but at present their high price limits 
their use. It may be, however, that the price of vanadium will 
be lowered when the alloy can be produced in the electric furnace, 
but due to the formation of carbides special attention must be 
paid to the making of a low carbon vanadium alloy. 

Deoxidation requires that during the whole process there 
should be a purely neutral or reducing atmosphere. Formerly 
this condition was only obtained in the graphite crucible, silicon 
being reduced from the crucible walls by carbon, and forming 
the reducing agent. For this reason crucible steels made from 


a pure charge were, up to the present time, the best obtainable, 
although low silicon material was only produced with very great 
difficulty because of the silicon reduced from the acid crucible 
walls. The melting of low silicon crucible steels had to be done, 
therefore, in costly alumina crucibles. 


(1) Ferro-manganese. For reasons of economy the ordinary 
blast furnace product is used with the average analysis 

Manganese . 80 . 00% 

Silicon 1.20% 

Phosphorus 0.25% 

Carbon 6 . 00% 

The rather high phosphorus content can be neglected, for only 
a small percentage of ferro-manganese is used so that the phos- 
phorus of the charge is practically not increased. Occasionally, 
pure manganese, which is naturally very expensive, is used for 
special purposes. 

(2) Ferro-chromium. Here also the cheap high carbon mate- 
rial is usually good enough, with the analysis 

especially when the alloy is added liquid. The carbon does not 
produce any gas, for the alloy is only added after the bath is 
deoxidized. Of course the carbon of the alloy must be consid- 
ered in figuring the carbon of the steel. The more expensive 
low carbon alloys are, however, used in many cases. 

(3) F err o- silicon. The best is the ordinary 50 per cent, elec- 
tric furnace grade. High silicon blast furnace pig irons can 
indeed be used, especially with such a market as the present, but 
the carbon of this grade of material is higher than desirable. 
In making high silicon steels, such as dynamo plates, etc., the 
90 per cent, alloy is used. 

(4) Lime. This should, of course, be as free from sulphur and 
phosphorus as possible. In burning the lime it should be par- 


ticularly remembered that with the use of high sulphur fuel, 
such as is generally employed, the lime takes up considerable 
sulphur, so that with large pieces of lime the sulphur is highest 
at the outside and decreases towards the centre. 

A nalyses 
Outer shell 



0.48% S 

Middle part 

... 21% S 

0.20% S 

Core . . 


0.06% S 

One is therefore bound to consider the use of raw limestone, 
especially for the formation of the refining slag. As lime free 
from sulphur is needed, the stone could be burned in a shaft or 
rotating furnace by means of waste gases, so far as they are 
available. Moreover, tests with the ring furnace have shown 
that the lime in a chamber does not show the same increase in 
sulphur at all parts of the chamber. An example is given below. 

The raw limestone used was very uniform and had 0.05% S. 
Tests taken from the material after being burnt showed the 
following results: 

Average test from the wall of the chamber o.n%S 

" in front of the fire o.i6%S 

' somewhat further from the fire o. 16% S 

" " at the door 0.09% S 

As lime burnt in the ring furnace is mostly used for other 
purposes, one is in the position to take the low sulphur part and 
use it specially. If burnt lime is bought it is well to consider 
the percentage of moisture and carbon-dioxide contained. On 
the other hand a small proportion of carbon-dioxide is not a 
great disadvantage to the process, for the slag must finally 
show a certain amount of carbon-dioxide, at least for desulphur- 
izing and deoxidizing. Also certain percentages of magnesia in 
the lime are a disadvantage as it makes the slag less fusible. 

(5) Fluor -Spar. This should be as low as possible in sulphur 
and phosphorus, and is suitably paid for according to its contents 
of fluorine. Also contained magnesia is a disadvantage. 

(6) Iron Ore for the Carbon Refining Process. Any ore, even 
brown iron ore, can be used but high percentage ore is recom- 


mended, so that the slag volume and the heat lost in the slag 
are not too great. Naturally, it is also better to use ore not 
too high in sulphur, especially if metal has to be worked that is 
high in carbon, sulphur, and phosphorus. Phosphorus in the 
ore, on the other hand, is not harmful so that minette ore can be 
used, for the bath cannot reduce phosphoric acid from the ore. 
(7) Carbon. A material should be chosen that is low in ash, 
sulphur, and volatile matter. Graphite, anthracite, petroleum, 
coke, etc., can be used according to one's wish and the market 
price. Below are given several analyses of these materials: 

Vol. matter 



Petroleum coke . ... 

7 e% 



Retort carbon 


I 8 

I 2 

Flake graphite 


6. 9 


They are best used in moderate sized pieces. If finely 
divided material must be used it is best to weigh it out into 
bags, or else briquette it, 


The advantages of melting in the electric furnace are chiefly 
brought about by the possibility of maintaining purely neutral 
or reducing atmospheres, which means that the hot materials 
do not attack the furnace walls as in the cupola, air furnace, etc. 
As is well known the melting of pig iron, etc., in the cupola 
is attended with a considerable absorption of sulphur, which 
sensibly affects the final quality. For instance: 

C% Si% Mn% S% 

Material before melting 3.50 2.90 1.20 0.035 

After melting once 3.40 2.70 0.055 

After melting twice 3.30 2.50 0.80 0.073 

Foundries making low sulphur material such as high quality 
castings, malleable iron castings, etc., must therefore melt either 
in the open hearth or air furnace, and notwithstanding this most 
expensive operation the undesirable action of the furnace gases 


on the charge is not prevented, as is shown by the following 
analyses : 

A . Melting of Pig Iron: 

C% Si% Mn% S% P% 

(1) Before melting 3.30 1.55 1.67 0.053 -36 

Finished material 3- 2 5 0.66 0.76 0.083 -37 

(2) Before melting 3.18 0.59 1.79 0.075 O- 2 7 

Finished material 3.16 0.22 1.22 0.093 O- 2 7 

(3) Before melting 3.06 0.72 1.98 0.069 O- 2 3 

Finished material 3. 02 0.28 0.28 0.090 0.23 

The melting in all these cases took place in an air furnace, 
using bituminous coal with about 0.7% sulphur. 

B. Steel Castings Melted in a $-Ton Open Hearth Furnace. 

c s 

(1) Before melting 0.050 

Finished material o . 30 o . 060 

(2) Before melting o . 037 

Finished material 0.25 o . 050 

(3) Before melting o . 048 

Finished material o . 45 o . 062 

Naturally those plants suffer which have to use, anyhow, 
high sulphur pig iron and fuel; on the other hand, with melting 
in the electric furnace there is no oxidation of the iron nor of the 
valuable constituents silicon, manganese, etc., such as is shown 
in the above analyses. As is well known in the melting of an 
ordinary foundry iron, a loss of at least 10% of the silicon is 
calculated, with higher silicon irons still more, and on this 
account less scrap can be melted than the silicon of the cold 
foundry iron would allow. In electric furnace melting there is 
no loss of iron, nor metal loss in the slag, for no melting slag is 

Electric melting is particularly important in the production 
of hard castings. The high manganese pig iron used suffers a 
high loss of manganese in melting in the ordinary furnace, which 
is entirely absent with electric melting, so that the amount of the 
expensive high manganese iron necessary can be considerably 
lowered. An important point that recommends the electric 


melting of iron for foundry purposes is that in the electric furnace 
the temperature can be governed as desired. The percentage 
of bad castings therefore ought to be somewhat reduced, for the 
casting of cold iron, which may happen with the cupola even 
with the most careful supervision of the operation, is excluded. 
Also in regard to the quality of the castings, electric melting 
should bring about considerable improvement, apart from the 
avoidance of an increase in sulphur, especially for pipe and thin 
walled castings, for which high phosphorus brittle material has 
now to be used in order to fill out the moulds. As it is possible 
to increase the temperature of a low phosphorus iron in the 
electric furnace to such an extent that the same fluidity is pro- 
duced as with a high phosphorus iron, one can therefore, under 
these conditions, produce thin walled castings from low phos- 
phorus iron which is not brittle, without getting porous or blow- 
holy castings. 

The electric furnace is also very suitable for melting ferro- 
manganese, and all the ferro alloys, which are so much used in 
steel plants and also recently in foundries. Every metallurgist 
knows that for quickly completing heats of steel in the furnace 
or in the ladle, considerably more ferro-manganese must be used 
if it is added cold than if added liquid. In spite of this, up to 
now, he has been forced to be satisfied with the use of solid pre- 
heated ferro-manganese, because the metallurgical furnaces 
available for melting this easily oxidizable material are not prac- 
tical as the loss increases immeasurably. The electric furnace is 
here particularly applicable, for with a reducing atmosphere an 
oxidation of the manganese is excluded. Arc furnaces will not 
be so suitable for this purpose, for on account of the overheating 
of the bath near the electrodes the manganese will vaporize, and 
at the same time the furnace lining will be strongly attacked 
by the manganese vapor. The introduction of the electric 
furnace into foundries, steel works, etc., will be further favored 
by the fact that during the melting an excellent mixing of the 
charge will take place. Until now in cupola melting one is 
compelled to take the metal more or less as it comes, even when 
making a special material, because though one may know the 


composition of the material charged it is difficult to figure on 
the loss during melting, and therefore on the final composition. 
With the electric furnace on the other hand, where there is no 
oxidation, one can calculate exactly beforehand the composition 
of the final fluid metal, apart from the fact that an absolutely 
uniform material, free from impurities, will be produced. Bad 
heats, because of low or high silicon, will be excluded because 
one can add to the bath the right amount of ferro-silicon on the 
one hand, or low silicon pig on the other. Attempts have often 
been made previously to increase the silicon in a low silicon iron 
by the addition of ferro-silicon to the casting ladle, a process that 
is only partially successful, for, to absorb the silicon, it must be 
first melted, which requires a very hot bath of metal and also a 
certain amount of time. Both conditions are fully met in the 
electric furnace, but not in the casting ladle. 

Also cast-iron scrap, turnings, etc., can be melted without 
the scrap being for the most part burned and slagged off as in 
the cupola. Indeed, this great loss, when melting fine material 
such as turnings, etc., in the cupola, has forced those plants 
which have considerable amounts of such scrap to briquette it 
before melting. The considerable cost of this process is always 
lower than the saving due to the decreased loss. Also, a low 
carbon material, similar to cold-blast iron, can be produced 
without difficulty by the melting in of wrought-iron scrap. 

The melting of pig iron in the electric furnace can, at the 
same time, be combined with a refining of sulphur or silicon. 
In regard to the sulphur its removal is easy if a lime slag is pro- 
duced. As the silicon content of the pig-iron is almost always 
high enough, desulphurization readily takes place according to 
the equations given before, if there is temperature enough. The 
sulphur enters the slag, which is drawn off. Such desulphurized 
iron is particularly suitable for malleable iron castings, so that 
even for this high grade material a cheap high sulphur iron or 
scrap can be used. The lowering of the silicon will naturally 
be brought about by the addition of a correspondingly low silicon 

As a conclusion it may be taken that the electric furnace will 


find a place in foundries, etc., for melting pig iron, ferro alloys, 
etc.; particularly in the production of high quality material. 
The advantages are an absence of loss by melting and in the 
slag, the production of the same metal as calculated theoretically, 
the use of cheap high sulphur iron, and the melting of more cast- 
iron scrap. In general the induction furnace will be preferable 
for these purposes, for the bath is heated uniformly and there is 
no electrode action on the metal which would give loss of manga- 
nese because of vaporization. 

The profitable use of the induction furnace in the melting of 
fine scrap must be particularly mentioned, for there is always a 
bath of metal in the furnace. On charging the cold scrap it 
immediately falls into this bath and is therefore protected from 
oxidation. The power consumption per metric ton of steel 
scrap melted in this type of furnace is about 580 kw. hrs., an 
amount that makes electric melting appear quite economical, 
when it is considered that the great loss present with any other 
kind of melting is entirely absent. 


Most large steel works that have several blast furnaces, as 
well as foundries taking metal direct from the blast furnaces, 
already have mixer plants either to regulate the production, to 
get a better mixture of the different casts, or to obtain as thorough 
prerefining as possible. This means a separation of the sulphur 
brought about by a part of the sulphur slagging off as a sulphide 
of manganese, if there is sufficient manganese in the iron. The 
size of the mixers varies a great deal from 25 to 1,000 tons and 
more capacity. The small mixers are preferably used for iron 
foundries such as pipe foundries, that take direct metal, but 
the prevention of cooling with the small mixers is naturally not 
very good so that sometimes heating is necessary. If refining 
is desired in the mixer, then heating by means of fuel is not so 
profitable and there is opportunity for the electrically heated 
mixer, which will be similar to an ordinary electric furnace of 
very large size. Furnaces of more than 25 metric tons capacity 
have, however, not yet been built, so that in this respect there 


has been no experience. The requirement for an electric mixer 
is that the metal should be held at the right temperature. The 
question should be best solved by an induction furnace, for here 
the temperature can be kept at any desired degree, and there 
is also certainty of an absolute uniformity of the whole metal 
because of the movement of the bath. In the mixer there would 
also be a thorough desulphurization of the metal, so that the 
product would undoubtedly meet the most rigid requirements of 

On the other hand, if the mixer metal is to be used for steel 
making, and then subjected to subsequent refining processes, 
heating with ordinary fuels would still in most cases be the 
more economical. 

Further, here again the known calculations give weight on 
one side or the other, namely, which is the more expensive under 
the conditions present, heating with electricity or direct heating 
with fuel. 


The refining of pig iron can be carried out very well in the 
electric furnace, and just as well by the oxygen of ores as by that 
of the air. In general the induction furnace here also would 
come mainly into consideration, for, as mentioned above, the 
refining process can only be carried out in the arc furnace with 
great electrode loss, and use of considerable refining material. 

The iron could be melted direct in the electric furnace, or the 
liquid metal could be charged from the blast furnace, mixer, 
cupola, open hearth or special furnaces, after the liquid metal 
had been previously refined, that is desulphurized, etc. 

Refining with ore in the electric furnace is, however, expensive 
because the reduction of the ore takes place slowly, exactly as in 
the Talbot, Bertrand-Thiel, and other processes, so that the 
current consumption caused by radiation is too great. In order 
to accelerate this reduction, and so save electric energy, one 
could consider charging the ore highly heated, if peat or some 
other fuel not suitable for steel works furnace operation is readily 
accessible. Concerning the thermal advantages brought about 


by the use of highly heated ore, the following approximate rough 
calculations give some information. 

The pig iron contains 3% carbon, and will be refined by pure 
magnetite heated to 800 C. According to the equation: 

232 kg. Fe 3 O 4 + 48 kg. C = 1 68 kg. Fe + 112 kg. CO, 

the 30 kg. carbon that are in a metric ton of iron require ^~ 


= 145 kg. ore. This amount of ore heated to 800 C. holds 

I ^OO 

145 X 0.2 X 800 = 23200 cals., which equals ~ = 26.8 


kw. hrs., an amount that helps the electric furnace considerably, 
so that the use of preheated ore is worth consideration, if the 
cost of preheating is not too high. 

There are also proposals to carry out air-blast refining, similar 
to the Bessemer, in the electric furnace. With arc furnaces the 
electrodes would have to be drawn up high during the blowing, 
so that during this operation no heat would be supplied, and the 
bath would chill, if there w T ere not sufficient silicon and phos- 
phorus present to balance the heat lost, and bring the metal to 
the casting temperature of soft steel. In this case the bath 
must be alternately electrically heated, then blown for a short 
time, but this gives so many operating troubles that the intended 
saving due to time saved with blowing is not realized. 

Iron low in silicon and phosphorus, that cannot be handled 
by either the acid or basic Bessemer, may be refined with a blast 
of air in the induction furnace, for here the bath can be heated 
during the blow. The following rough calculations give some 
information on the probable results with a lo-ton furnace and 
a pig iron with 3% carbon, and a temperature of 1300 C. 

10 tons iron contain 300 kg. carbon, which would require 

(12 C + 1 60 = 28 CO) - - = 400 kg. 0, 100 kg. air con- 

tain 23 kg. 0, so that 400 kg. correspond to - 

1740 kg. air. The temperature of the bath must be raised from 

1300 to 1650 C., that is 35oC. 

The blast may leave the bath at an average temperature of 


1500 C., although this value is probably too small, for the 
carbon in the bath heated to 1300 C. will burn at a high tem- 
perature, and it appears doubtful whether the very hot gas 
produced will give its heat to the bath completely enough to 
escape at only 1500 C. 

There is therefore the following amount of electric energy 
conducted to the bath. 

10000 kg. iron heated 350 C. 10000 X 0.2 X 350 = 700,000 
1740 kg. air heat to 1500 C. 1740 X 0.3 X 1500 = 783,000 

Brought in: 

300 kg. C burnt to CO ...................... 741,900 

Leaving ..................................... 741,000 

This corresponds to T- - = 857 kw. hrs. 

Therefore an electric induction furnace of 10 tons capacity 
will operate with about 800 to 900 kw. hrs. If the efficiency of 
the furnace is taken as 60%, then 480 to 540 kw. will be sufficient 
to heat the bath. This shows that the carbon can be thoroughly 

removed, in - to , that is i^-i^ hours. The time interval 
480 540' 

under these conditions compared with that of the other air-blast 
refining processes is very considerable. Air-blast refining must 
therefore be carried out extremely slowly, or, with frequent 
interruptions. The proposals to refine in this way have so far 
found no practical application. 

From all this it follows that, if there is a mixer available 
it is a good thing in all cases to bring about as complete refining 
as possible in the mixer, and so relieve the electric furnace. 
Better economic results are obtained in refining pig iron if large 
amounts of mild steel scrap are available to melt with it, so that 
only a hard steel-like product remains to be treated. 

The problem of pig-iron refining in the electric furnace now 
approaches solution, because, for example, for railroad material 
there is an inclination to use harder qualities of steel than are 


used now. If, in a case like this, the pig iron to be used is 
sufficiently low in phosphorus so that no refining is necessary, 
to give a steel low enough to meet specifications, then carbon 
alone has to be removed. In such a case the electric furnace 
could work economically in many places. Also with specifica- 
tions calling for very low phosphorus in the steel the bath high 
in carbon and phosphorus can be dephosphorized without re- 
moving the carbon, as mentioned above, by keeping the tem- 
perature low and forming an easily fusible basic slag, containing 
oxide of iron, which is drawn off. when the phosphorus is low 
enough in the steel. 

The refining of pig iron in the electric furnace is not advan- 
tageous if a low carbon, absolutely phosphorus free, material 
has to be made from a high phosphorus pig iron. In order to 
remove the phosphorus the carbon must first be completely 
taken away, and the bath even overrefined to a certain extent. 
In general for this purpose the electric furnace cannot compete 
economically with the open hearth. Further, in this case the 
cost of fuel and of current have to be weighed against each 

If the pig iron to be refined is high in silicon, as well as 
phosphorus, two electric furnaces can be used, one with an acid 
lining for removing silicon and carbon, the other with a basic 
lining to remove the phosphorus. Still such an iron can be worked 
in the basic furnace in which case a sufficient amount of lime 
must be added to prevent the lining from being attacked. Fi- 
nally, the deoxidation, etc., can be carried out in a third furnace 
with an acid lining, or in crucibles, and many combinations of 
the crucible, open hearth furnace, mixer, converter, etc., are 
possible, the suitability of which must be decided for each 
separate case. 

The output when refining with ore is extraordinarily high, 
because there is no loss and because of reduction from the ore, so 
that it is over 100%. 



High quality steel production aims at the melting of the 
softest to the hardest qualities as desired in both alloy and plain 
steels. Steels that will meet the most rigid requirements in 
regard to low sulphur and phosphorus on the one hand, and on 
the other hand be as free from oxygen as possible and in this 
respect be equal to crucible steel. 

The best material up to the present has been made by the 
crucible process, special care being paid to the kind of raw 
material charged. The \ise of the purest materials is a first 
requirement for the crucible process, for naturally no removal 
of sulphur and phosphorus is possible to any considerable extent. 
Indeed this dependence on certain kinds of iron, which meet 
the guarantee of absolute purity, altogether apart from the 
cost, has finally been the reason for the introduction of the 
electric furnace, as it made the material forming the charge 
independent of a fixed source of supply. The general strike in 
Sweden during 1911 has opened the eyes of the leaders in this 
industry to the disadvantages that may come when one is 
forced to use a certain material alone. 

In comparison with this the electric furnace offers the great 
advantage that it is not dependent on any certain special raw 
material, for the most impure materials can be refined so that 
they become even better than the purest Swedish charcoal iron 
in regard to purity from phosphorus and sulphur. At the same 
time deoxidation takes place just as completely as in the crucible 
because a purely reducing atmosphere is maintained, the steel 
can be held as long as desired, and the temperature can be regula- 
ted with more certainty than in furnaces heated with fuel. Sim- 
ilar conditions are not offered by any other metallurgical furnace, 
for in them the action of the flame on the bath cannot be entirely 

From all this it is seen that electric steel is at least of equal 
value to crucible steel, for it can be produced practically free 
from phosphorus, sulphur, and non-metallic inclusions. There- 


fore the use of the electric furnace gives the advantage, that in 
it ordinary low carbon steel can be improved and made equal 
to the very best qualities of crucible steel, for from the low carbon 
steel : 

(1) Phosphorus and sulphur are completely removed. 

(2) It is totally deoxidized. 

(3) It is freed from slags and inclusions. 

(4) It is accomplished at a lesser cost per ton. 


It is a side issue in what way the steel is prerefined, whether 
in the converter, in the basic or acid open hearth, or in any 
other way. Also the material can either be charged liquid or 
cold, but in the latter case, the electric furnace will also be used 
for melting. If the electric furnace is worked in combination 
with an ordinary steel plant from which it obtains its charge, 
then it is most suitable to pour a part of the steel works charge in- 
to the electric furnace before the deoxidizing additions are made. 
On the other hand, if larger heats are made in the steel works 
than the electric furnace is able to take, or for other reasons, then 
the bath of steel to which the additions have already been made 
can be partially poured into the electric furnace. Naturally 
it is preferable that the electric furnace be able to take the whole 
heat with the restriction that the building of very large electric 
furnaces is at present troublesome. In regard to this the first 
part of the book should be consulted. 

In the case we are considering, the aim of the electric furnace 
is to improve the steel and to produce a material of equal value 
to open hearth, or crucible steel, in particular : 

(1) To recarburize the bath to the required hardness. 

(2) To deoxidize the bath. 

(3) To bring about the removal of gas and slag inclusions. 

(4) To alloy the bath as desired. 

Such a process can be profitable, for example, if there is a 
Bessemer plant operating, and it is desired to produce from the 


Bessemer steel material equal to high grade open hearth, which 
will be required for various purposes such as boiler plate, etc. 
Such material finished by the electric furnace is extremely 
suitable for particular purposes, so much the more that it rolls 
and forges well and shows considerably increased physical 

If material is charged into the electric furnace to which no 
deoxidizing additions have been made, then it has to be first 
completely deoxidized, and the production of steel of a satis- 
factory quality requires, in the first place, that this deoxidation 
be carried out very carefully. As shown in a previous chapter 
it can be done in many ways, with ferro-manganese, ferro-silicon, 
etc. Which of these materials should be used depends on the 
quality of steel that has to be produced, particularly whether 
it is to be a low or high manganese. 

If a product is to be made as low in manganese as possible 
then it is best to carry out the deoxidation with ferro-silicon. 
Immediately after pouring the charge into the electric furnace 
the first addition of ferro-silicon should be made, preferably in 
pieces about the size of one's fist, and at the same time the 
bath should be covered with an easily fusible slag to exclude 
the air. 

The kind of slag is governed by the furnace lining: with a 
basic hearth a neutral or basic slag is charged ; with a neutral or 
acid hearth, on the other hand, one of greater acidity. A suit- 
able mixture of lime and sand with more or less fluor-spar may be 
used to form the slag, all of proper size. In regard to the amount 
of slag it should be mentioned that the bath has only to be 
completely covered. Furnaces of greater capacity that work 
with a deep bath use, therefore, a lower percentage of slag than 
those of less capacity which expose more surface per ton. 

The first addition of ferro-silicon is given, before the bath 
is covered with slag in order to save time in the first place, 
and secondly to prevent the light ferro-silicon from being en- 
closed in the slag, which is quite thick at first. The slag first 
turns black due to the absorption of oxide of iron from the bath, 
for a condition of equilibrium is formed between the oxide dis- 


solved in the bath and the slag. For complete deoxidation it is 
therefore absolutely necessary that the slag contain no oxide, 
and so it must always be snow white. This is produced by 
sprinkling a suitable reducing agent on the slag when it shows a 
dark color, and maintaining a neutral or reducing atmosphere. 
In the induction furnace fine ferro-silicon, about pea size, is 
used in this way, being added from time to time in small amounts. 
In the arc furnace this ferro-silicon can be produced from the 
slag if carbon is added. Which of the two methods is the cheaper 
we will not investigate. 

If the slag keeps snow white then the melter takes pouring 
tests and convinces himself of the condition of the steel, and if it 
does not yet pour quietly, adds more pieces of ferro-silicon to 
the bath until a further test gives a good result. A completely 
deoxidized steel, melted with a white slag, must pour without 
trouble. Alloys such as nickel, manganese, chromium, etc., 
can now be added in the theoretical amounts, for no slagging 
of these additions can take place under the white slag covering. 
No preheating is necessary, and also the cheaper ferro alloys, 
high in carbon, can be used in the induction furnace for, due 
to the movement of the metal bath in this furnace, it is impossible 
for carbides to remain undissolved. 

In making a steel with low to average manganese content it is 
the best to give first an addition of ferro-silicon, after which the 
bath is covered with slag. The final deoxidation can now be 
made with ferro-manganese, spiegel, etc. After this the slag will 
first darken, due to the manganese reacting with the oxide of 
iron forming MnO, which enters the slag. As mentioned above 
the first requirement for complete deoxidation is that the slag 
be snow white. The black slag produced must, therefore, be 
reduced in a suitable manner and made white. 

In the induction furnace ferro-silicon again serves as a 
suitable reducing agent and carbon in the arc furnace, both of 
which are sprinkled on the slag. The reduced manganese goes 
again into the slag. It therefore passes through a cycle and 
really only serves as a bearer of the oxygen contained in the 
metal, so that the deoxidation can be carried out with very small 


amounts of manganese, and a final material with a moderate 
percentage of manganese can be produced. 

Instead of ferro-manganese, manganese ore can be used, 
and in the induction furnace this ore is selectively reduced by 
ferro-silicon rather than with carbon, but in the arc furnace on 
the other hand the one would be reduced by carbon under the 
influence of the arc. It is not necessary to consider here whether 
it is the cheaper to use ferro-manganese melted in the blast 
furnace or to reduce manganese in the electric furnace from 
manganese ore. The remainder of the process of melting is 
the same as that described under the production of steel free from 

If the material has to be harder, that is higher in carbon 
than the material charged, then after the first addition of ferro- 
silicon before the slag is made, the necessary amount of carbon 
is added to the bath. A small excess is given, depending on the 
size and the physical condition of the carbonizing material used, 
for a part is burned as it is charged into the furnace. Then 
comes the slag formation, and it is well to take a test and make 
a quick color carbon determination to see whether the metal is 
of the right hardness. The further process is the same as that 
used for the production of low manganese steel. The tempera- 
ture is held at such a degree that the small impurities caused by 
the reduction can separate readily and is gradually increased 
to the proper casting temperature. It must be remembered 
that the slag formation requires a certain amount of heat as 
also the solution of carbon, if any is added, and so cools the 

At the beginning of the deoxidation it is well to give an 
addition of ferro-silicon even when melting high carbon material. 
If carbon is added to the bath before the ferro-silicon then there 
is a vigorous action, and a considerable loss of carbon cannot 
be avoided. At the most, therefore, one can only add a part 
of the carbon before the ferro-silicon, and after the deoxidaticn 
the rest must be added in a special operation to give the hardness 
required. Also by the addition of carbon before the ferro- 
silicon scarcely any ferro-silicon will be saved, and on the other 


hand it only increases the time of the operation and the work 
in the furnace. 

If the ordinary forging and pouring tests are favorable then 
the casting of the heat is proceeded with. The slag also is 
poured into the casting ladle in order to protect the metal from 
the influence of the air in the first place, and secondly to prevent 
the slag sticking to the hearth of the empty furnace, and attack- 
ing the lining. Due to the total deoxidation the electric steel 
casts so quietly that the addition of aluminum to the stream 
during pouring is absolutely unnecessary, so that in this respect 
the quality of the metal does not suffer. Casting is carried out 
in just the same way as in plants making high quality steels. 

As a conclusion it may be mentioned that during this finishing 
process considerable desulphurization is brought about by the 
silicon present, even when it is not intended, so that in this 
respect the after treatment of low carbon steel in the electric 
furnace means a rather considerable improvement in quality. 

If an addition of ferro-manganese for deoxidation has already 
been made in the converter, open hearth, etc., then after pouring 
into the electric furnace an easily fusible basic slag alone has 
to be made and kept constantly white, that is free from oxide. 
This is brought about as mentioned above by ferro-silicon in 
the induction furnace or just as well by carbon in the arc furnace, 
where silicon is reduced from the slag by the influence of the 
arc. This assumes that the steel poured into the electric furnace 
already contains the necessary percentage of silicon, but if this 
is not the case then before the formation of the slag the corre- 
sponding addition of ferro-silicon is given. Naturally in this 
process also there is a lowering of the sulphur of the charge, 
even if such is not intended. 


Here also it does not matter in what way the steel is pre- 
refined, or whether it is charged hot or cold so that the electric 
furnace has to be also used as a melting furnace. The aim of 


the after treatment in the electric furnace is to raise the quality 
either to that of open hearth or the best crucible steel. The 
metallurgical process in the electric furnace must therefore im- 
prove the steel in regard to the following points: 

(1) Eliminate the phosphorus. 

(2) Deoxidize and desulphurize. 

(3) Remove the gas and slag inclusions. 

(4) Recarburize or alloy according to requirements. 

In regard to the removal of phosphorus and sulphur both 
elements cannot be removed from the bath in one operation, for 
the removal of the phosphorus takes place by an oxidizing or 
refining process, and that of the sulphur, on the other hand, by 
a reducing process : 

FeS + CaO + C = Fe + CaS + CaO. 

These operations must, therefore, be carried out one after 
the other, and it is similar in principle whether the bath is de- 
sulphurized first and then dephosphorized, or, on the other hand, 
dephosphorized first and then desulphurized. Which of the 
two ways is the most suitable depends on the composition of 
the charge that is put into the electric furnace and on the kind 
of steel to be produced. 

If the metal as charged has had no additions so that it is 
not yet deoxidized, then it is best to dephosphorize first of all 
as the necessary conditions are present. The total removal of 
the phosphorus requires an overoxidation of the bath, so that 
the metal gives a seamy, that is a red short, forging test. When 
using a charge that is still oxidized therefore only a basic slag 
has to be charged to favor the soaking of the bath with oxygen, 
and at the same time give conditions so that the phosphoric 
acid formed is immediately combined with lime. The require- 
ments in regard to freedom from sulphur on the part of the slag- 
making materials are not particularly high, for the bath has to be 
desulphurized afterwards anyhow, but it is well to use materials 
as low in sulphur as possible, so as not to raise the sulphur in the 
bath. Also the phosphorus percentage of the ore is without 
influence, for an oxidized bath cannot reduce phosphorus, and 


is therefore unable to take phosphorus from the ore or lime- 

If it is desired, on the other hand, to desulphurize first then 
the bath must be first completely deoxidized, and after the re- 
moval of sulphur the bath must be oxidized again to remove the 
phosphorus. The whole manipulation of the deoxidizing, there- 
fore, gives no lasting result, and at the same time only the purest 
ore and limestone can be used to form the refining slag so as to 
prevent absorption of sulphur by the bath. 

On the other hand if the charge has already had an addition 
of ferro-manganese in the converter, open hearth, etc. ; and if a 
very soft steel has to be made with, the lowest possible phos- 
phorus and sulphur and also practically free from silicon, then 
it is best to desulphurize first. In this way one can work with 
a small amount of silicon, and the low silicon remaining in the 
bath is removed during dephosphorizing. The heat is then 
finished with the addition of ferro-manganese, and is cast just 
as in the open hearth process, keeping the slag back. Here also 
it is necessary to use ore and lime free from sulphur to prevent 
the bath again taking up sulphur during the subsequent opera- 
tions. The quality produced is, however, only equal to that of 
good open hearth not crucible steel. If sulphur-free burnt 
lime is not available then it is well to use raw limestone if the 
sulphur is low enough. Apart from this and some other special 
cases the charge, prerefined and then poured into the electric 
furnace, will always be dephosphorized first whether deoxidizing 
additions have been already given in the preliminary furnace 
or not. This is done because: (i) With desulphurizing first 
the silicon used is again removed during the dephosphorizing. 
Therefore the amount of ferro-silicon used is unnecessarily in- 
creased; (2) the total deoxidation, which is necessary for de- 
sulphurization, would be made of no use by the subsequent 
oxidation, (3) the ore and limestone used for the refining slag 
have to be very free from sulphur. 

Immediately after pouring the charge into the electric 
furnace a refining slag should be formed with ore and lime. 
The size of these materials should not be too great as otherwise 


they only fuse together with difficulty, that is, the formation 
of the slag takes too long and the time of the heat is increased. 
On the other hand there is no limit to the fineness of the material, 
so tKat, for example, unbriquetted concentrates can be used. 
The burnt lime is best broken up just before charging, for it 
quickly takes up moisture and carbon-dioxide from the air. 

The amount of slag necessary is proportionally small, espe- 
cially if dephosphorization has already taken place to some 
extent, for instance, to o. i%. The bath need not be well covered 
by the slag, although dephosphorization naturally takes place 
more quickly if the slag covering is not too small. On the other 
hand it is well not to unnecessarily increase the amount of slag 
so as to avoid loss of heat. In the induction furnace, work in 
a high manganese charge, that is one already deoxidized in the 
first furnace, it is well to make a slag with i% ore and 2% lime 
of the weight of the charge. If the forging test shows that the 
bath has the right percentage of phosphorus, then the slag is 
drawn off, and the last traces removed by means of fresh lime 
thrown over the bath. This thickens the remainder of the slag 
so that it can be easily removed. 

If a high phosphorus charge is to be worked, then it is well 
not to charge the whole amount of slag necessary at one time for, 
as mentioned before, it is not recommended to work with too 
large a slag volume in the electric furnace. In this case it is 
better to charge the ordinary small amount of slag, remove it 
when completely used up, and then form a new slag of the same 
weight. This should be repeated as required. Such a case 
can happen in practise if the electric furnace charge is taken 
from a heated prerefining furnace, such as a Wellman-Talbot 
furnace, in which by exceeding the capacity a material is produced 
that is high in phosphorus. 

If the electric furnace is to be used continuously for the 
refining of such high phosphorus charges, then it is well to figure 
on this in the construction of the furnace. A shallow bath but 
a large surface should be used in order to give the refining slag a 
large attacking surface, and to shorten the time of the process. 
On the other hand the surface of the bath should not be too great 


for, in this way, the radiation loss increases and the efficiency 
of the furnace drops. These are important points for the de- 
signers of the furnace, namely, to make the hearth the right size 
and shape to properly meet the conditions, and in this respect 
experience obtained with open hearth furnaces will be valuable. 
If a high phosphorus charge is worked, with several refining 
slags, then the first slag is very low in iron but high in lime and 
phosphoric acid. These slags are of some value in agriculture 
as low phosphate slags, so that they need not be thrown away. 
The succeeding slags, however, are rich in oxide of iron and low 
in phosphorus, and can be used as a first slag for subsequent 
heats and so be used more completely. If the fluxes are used 
to exhaustion then the consumption of ore and lime, in reference 
to the finished material, is considerably reduced. 

The bath can now be deoxidized, carburized, and desul- 
phurized. The removal of the oxygen and sulphur takes place 
together, the first by means of ferro-silicon, carbon, or ferro- 
manganese depending on the kind of material to be melted; the 
latter by the addition of ferro-silicon and in the arc furnace by 
silicon or calcium reduced from the slag by carbon under the 
influence of the arc. 

First an addition of ferro-silicon is made that can be a little 
more in amount than is necessary for deoxidation alone because 
of the desulphurization also taking place. In general the opera- 
tion is exactly the same as described in the previous section, 
namely: "The production of electric furnace material from 
previously refined metal with low phosphorus and sulphur." 

It may be mentioned that tungsten has similar desulphuriz- 
ing properties to silicon, so that tungsten heats can also be made 
extremely low in sulphur. 

A peculiar phenomenon must be mentioned which can 
happen under certain conditions with non-expert handling of 
the electric furnace. As already said a well deoxidized charge 
pours quietly, and only pipes a little on solidifying, depending 
on the temperature and the silicon. On the other hand, if the 
heat is made too hot, if the bath is not completely covered with 
slag, if air can enter the furnace, or if the deoxidation slag is 


not kept sufficiently free from metal, then the steel casts very 
badly, even though it may contain several tenths per cent, of 
silicon, the forging tests will show the properties of an oxidized 
material. The causes for this phenomenon are not at present 
very clear, but it appears probable that at high temperatures a 
part of the silicon occurs dissolved in the metal as a suboxide, 
probably with the formula SiO. Because of the similarity 
between silicon and carbon the possibility of an alloy of iron and 
silicon-suboxide can be thought of, for, as mentioned before, the 
existence of an alloy of iron with carbon-monoxide is probable. 



The course of the metallurgical reactions in the Heroult 
and the Rochling-Rodenhauser furnace is given in the two 
accompanying diagrams (page 398). The curves for the Her- 
oult furnace were published by Thallner in No. 5, 1909, of 
Kohle und Erz, and are taken from a heat in a 3-ton furnace; 
while the diagram of the Rochling-Rodenhauser furnace is taken 
from an ordinary heat made at Volklingen in an 8-ton alternat- 
ing current furnace built up to take 5 to 6 tons. 

From a comparison of the two diagrams it is seen first that 
the time of heat in the Heroult was twenty minutes longer than 
in the Rochling-Rodenhauser furnace, notwithstanding that the 
former was only worked with a 3-ton heat, while the latter had 5 

Also, the material made in the Rochling-Rodenhauser is at 
least just as pure as that produced in the Heroult, notwithstand- 
ing that a much more impure charge was worked in the former. 
The oxidation period is distinguished, in the Rochling-Roden- 
hauser furnace, especially at the beginning, by an extraordinarily 
quick removal of the phosphorus and manganese from the 
steel. For instance, in the first twenty minutes the phosphorus 
drops from 0.06 to 0.025, that is 0.035%, while in the Heroult 
furnace it is only lowered from 0.03 to 0.02, that is 0.01%, in the 
same time. 

Also the manganese drops from 0.49 to 0.12, that is 0.37% in 


the first hour in the Rochling-Rodenhauser furnace, while in the 
Heroult furnace in the same time it is only lowered from 0.21 to 
is 0.1%. 

From this it follows that the Rochling-Rodenhauser furnace 
must be considered as a good oxidizing furnace. The lowering 
of the phosphorus content of the slag, which is to be noticed 
during the oxidation period with both furnaces, is due to the 
slag being diluted from time to time by the addition of roll scale. 

The removal of sulphur during the oxidation period is rela- 
tively unimportant with both furnaces, and only after the 
recarburization, or after the formation of the final slag, does the 
real desulphurization begin. During this period the sulphur 
is lowered in the Heroult furnace from 0.07 to 0.012%, while 
in the Rochling-Rodenhauser a desulphurization from 0.065 
to traces is brought about. That the ability of the slag in the 
Rochling-Rodenhauser furnace to absorb sulphur is at least 
as great as that in the Heroult furnace is seen from the sulphur 
content of the slag, which is 1.25% in the first case, and only 
about 0.06% in the latter as shown by the curves. The amounts 
of slag -making constituents used in both cases are shown in the 
diagrams, so that all the details of the refining operation are 
given that are of interest. 


As already shown any material from the mildest to the 
hardest quality can be made in the electric furnace. Electric 
furnace material is distinguished by its freedom from gas and 
slag inclusions, and can easily be produced with very low man- 
ganese and completely free from phosphorus and sulphur, and 
as soft and forgeable as the Swedish qualities. This low carbon 
electric steel can easily be alloyed, for instance with silicon, for 
making material for dynamo plates, etc., the production of 
which in the open hearth furnace is troublesome because of the 
necessary low casting temperature. The use of the electric 
furnace for this purpose means, therefore, considerably easier 
operation. Also the softest material can be flattened out very 
thin without showing red shortness, and can be used for punching 


Slagging, recarburizing 
making the second slag 
Oxidation period' Reducing period 



















Reducing period 






j <L 













* ~> 











/ , 



' 1 







' s/ 

' / r 





m t 


Pji^ Sample-from the Ingot g III 





Mn0 2 


- t 





ffi r ' ' ' ' 
-T st^gKfe&ssss 

Scale for the lightly drawn Curves ?e F e O MnO SfOo CaO MflrO in the Slag 


, ,T, , , ,T, , , , 1 , , , ,Ti i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 n M i ii 1 1 1 1 1 1 1 n 1 1 1 it i i M 1 1 1 1 ii i iTi 1 1 1 1 1 1 1 1 


~ J 1& 

cale for the heavily drawn Curve.8 P S Mrr C -Si in the bath P ami S in the Sl 

s P s ? l..fiJ..lJ.J,J,J,,lJ,J.J,J,J,J tl ,LLL 



: s 








1 i 




/ j 



r 5 05" 

- r 



- -4 

--3 : 




1 | | 

-rr T 1 

r y 


Sample No 







'Ti^DXa i'P 

it 'in 

J Minutes 

j. Sample No. 


| ._.. 


15 45 

5 45 

2 30 3 



\\ i - i i 1 f i i 

234567 y 10 n 


2 34 50 7 80 10 

11 12 13 

Refining curves for the Rochling-Rodenhauser 

Refining curves for the Heroult furnace. 


and deep drawn work, deep stamping, etc., for everything where 
value is put on good malleability; also for the production of 
chains, the making of tools, etc., and further in those cases 
where especially soft open hearth qualities from Sweden must be 
used, such as seamless tubes, horseshoe nails, etc. On account 
of its purity this low carbon electric steel is much less inclined 
towards segregation than ordinary low carbon steel, and on this 
account should be particularly used where the highest require- 
ments of absolute certainty against brittleness are necessary. 

In the electric furnace construction steels of any degree of 
hardness can be produced, of any desired physical properties and 
chemical analysis, also alloyed with chromium and nickel where 
it is a question of meeting the highest specifications. These 
steels at present must be made in the crucible. 

With the large heats 1 possible in the electric furnace a cer- 
tainty of absolutely uniform composition is guaranteed for the 
finished steel, such as is suitable for the production of large 
forgings low in manganese. The electric furnace material can 
be easily hardened, and on account of its homogeneity and 
freedom from slag is an excellent material in such cases where 
the surface must be dense* and highly polished and show no 
cracks, such as running taps, etc. In general Plates 3 and 4 
show what high requirements are in every way satisfied by elec- 
tric steel. 


For the smelting of ore the electric hearth as well as the 
shaft furnace is to be considered, and in each particular case it 
must be carefully decided which type of furnace has the advan- 
tage. If very finely divided ores, high in sulphur, are to be worked 
up into steel by means of small sized reducing material, then the 
induction hearth furnace should be chosen because of the possi- 
bility of producing steel direct from such raw materials of any 
quality desired. Also because when changes have often to be 

1 See Osborne Amer. Electro-Chemical Society, 1911, Vol. XIX. 
* For quality of steel, see also Vom Baur, American Foundrymen's Associa- 
tion, May, 1911, page 247. 


made in the kind of metal produced, the making of valueless 
transition products is avoided. On the other hand, if coarse 
low sulphur lump ore and fuel are available, then the induction 
shaft furnace should be chosen, especially if the same quality of 
metal is always to be made, and the reducing material is high in 

The electro-thermal smelting of iron ores can naturally only be 
considered economically when the saving of coke, etc., compared 
with the ordinary blast furnace operation is greater than the ex- 
pense of the necessary electric power, so that it is dependent on 
the local prices of coke, etc., on the one hand, and electric energy 
on the other. Electric ore smelting will, however, be favored 
when one considers that considerably less capital is necessary 
for the plant than for the building of an ordinary blast furnace 
plant with the same output. Also the depreciation, etc., per 
metric ton of iron produced, are considerably lower than with 
the ordinary blast furnace. 

It must be further remembered that the quality of electric 
pig iron is higher than charcoal pig iron, and therefore it should 
command a higher selling price than the best charcoal iron. For 
steel making an iron can be readily made low in silicon, which 
only needs removal of carbon to make steel and forgeable metal. 
If this refining is carried out in the electric furnace, then it has 
to compete with the open hearth furnace. Recently Engelhardt 
at the meeting of the "Verein deutscher Ingenieure," in Berlin, 
made an interesting comparison between the open hearth furnace 
on the one hand, and different types of electric hearth furnaces 
on the other, namely the Heroult, the Girod, and the Induction 

For medium furnace sizes with these three types the produc- 
tion per h.p. day, with a cold charge, is taken as 20 kg. Cer- 
tainly this treats the induction furnace somewhat unfavorably, 
for it has about 10% greater efficiency. The given power con- 
sumption corresponds to 880 kw. hrs. per metric ton of steel. 
The electrode consumption, according to the most recent publica- 
tions, amounts to 28 kg. per metric ton in the Heroult furnace 
and 17 kg. in the Girod, while with the induction furnace, of 


course, there is not any. The consumption per metric ton of 
steel is therefore: 

With the Heroult furnace 880 kw. hrs. + 28 kg electrode. 

With the Girod furnace 880 kw. hrs. + 17 kg. electrode. 

With the Induction furnace 880 kw. hrs. + o kg. electrode. 

In the following table the results of calculations are given to 
show how much the kw. hr. ought to cost in order that the 
electric furnace may compete economically with the open hearth 
furnace using a certain amount of coal per ton at a certain price. 
The electrodes are taken at 26 marks per 100 kg. ($61.90 per 
metric ton): 





Cost of coal 


of coal per 

per metric ton 




by weight 

metric ton 







O. 1000 








o . 0476 




I. 7 8 

o . 0833 


o . 2023 













I. 7 8 







o . 0452 










o . 0690 









o. 1642 






o . 0428 















o . 2047 



It must, however, be remembered that even when producing 
an ordinary open hearth quality of steel in the electric furnace, 
material is produced with improved physical properties so that 
even with a continued regular output there should be a small 
increased price. If this increase in price is only 5%, and the 
price per metric ton of open hearth quality be taken as 140 marks 



Single Tube test piece from Plate I. 


($33.33), then there is a surplus per ton of 7 marks ($1.66), which 
with a power consumption of 880 kw. hrs. per ton makes almost 
0.2 cents that the price of power may be increased over the value 
given in the table. It may be further mentioned that the 
electrode consumption figures taken by Engelhardt appear 
somewhat too high, in view of the most recent figures given in 
Part I of this book, which vary from 10 to 15 kg. per metric 
ton of solid charge. 

Therefore the table (on page 281) may be referred to 
where the electrode consumption, however, is not considered 
at all; and where the heating costs alone are compared, on the 
one hand with the use of fuel and on the other with electricity. 
This table, therefore, gives results similar to those in the table 
on the preceding page for induction furnaces, with which it 
agrees exactly. 

The result of all this is that the electric furnace will not only 
play an important role in the future, but that it is already a 
factor which each iron and steel plant must now carefully con- 


Abbreviations used, xix 

Action of the electric current, 26-44 

Additions, 296, 376 

Advantage, chief, of electric furnace, 74 

Advantages of the electric furnace, 65-66, 74, 288-388 

Alternating current, comparison single and polyphase generators, 70 

current theory, 47-65 

polyphase current in general, 60 
Amperes, unit of measurement, 12 
Aluminum, 375, 392 
Applicability of the electric furnace (general), 73-74 

of the Girod furnace, 153 

of the Heroult furnace, 138 

of the Kjellin furnace, 188 

of the R. & R. furnace, 219 

of the Stassano furnace, 118 
Arc furnaces in general, 77-79 

heating, 37 

important points concerning, 105 

lengths with the Heroult furnace, 127 
with the Stassano furnace, 115 

temperature of, 78, 105 

the electric, 77-79 
Arrangement of an electric pig-iron furnace, 231-348 

of a Giro.d furnace, 145 

of a Heroult furnace, 123-128 

of a Kjellin furnace, 174 

of a Rochling-Rodenhauser furnace, 197-203 

of a Stassano furnace, no 

principle of induction furnaces, 164 

tilting furnaces in general, 72 
the Girod furnace, 152 
the Heroult furnace, 136 
the Kjellin furnace, 187 
the Rochling-Rodenhauser furnace, 215 
the Stassano, 117 
Arsenic, 372 

Auto-regulating transformer, 214 
Auxiliary apparatus, 300, 301 
Angular velocity (m = 2 x v), 48 

Daily billet-heating furnace, 253 


406 INDEX 

Basic bottom and lining material, 282 

bricks, 282 

Basic laws of electricity and magnetism, 11-25 
Blast furnace, total operating cost, comparison, 302 
Borchers, laboratory furnace, 33 

Bottom electrodes, influence of the Chapelet furnace, 243 
of the Girod furnace, 156 
of the Keller furnace, 226 
Bricks, basic. 282 

carbon, 281 

carborundum or silicon carbide, 307 

dinas, 281 

dolomite, 282 

half schamotte, 281 

half silica, 281 

magnesite, 282 

silica or acid, 281 

Canadian Commission, report of Haanel arid, 9 
Carbon, 372, 378 
bricks, 281 
electrodes, their efficiency, 84 

their influence with the Girod furnace, 155 
with the Heroult furnace, 135 
with the Stassano furnace, 113-119 
heat conductivity and specific resistance, 91 
mixtures for linings, 281 

necessary per ton of pig iron for desulphurizing, 338 
Castings, electric steel, 382 
Cast iron, low, phosphorus, for thin walled castings, 380 

melted in electric furnace, advantage, 380 
Chapelet's ore furnace, 242-243 
Charge, course of operations of electric furnace, 397 

nature of, 287 
Chemical action of the electric current in electric furnace, 37-38 

balance for ore smelting, 334 
Chrome iron ore, 282 
Clay used as a binder, 280 
Colby furnace, 169 

medal for, in consideration originality his furnace, 166 
Cold charge, melting in the Girod furnace, 159 
in Heroult furnace, 126-135 
in Kjellin furnace, 176-184, 192 
in Rochling-Rodenhauser furnace, 207-208, 220, 223 
in Stassano furnace, 112-113 
Combined arc and resistance furnaces, 79, 245 
resistances, 20 

arithmetical examples, 21, 22 

INDEX 407 

Comparison of costs, crucible and electric, 291 

open hearth and electric, 288 
Compressed air hammers, 285 
Conductivity, 12 
Conductor, resistance of a, 12 
Conductors, action of two on each other, 43 
between magnet and electric, 40-41 
of the second class, 15-17 
Construction of Girod furnace, 144 
of Heroult furnace, 124 
of Kjellin furnace, 173 
of R. & R. furnace, 197 
of Stassano furnace, 107 
Cooling of the electrodes, 99-102 
Copper, 372 

Cos <f> influence of the power factor, 55, 58, 64 
Cost of the auxiliary apparatus, 300 

comparative, ordinary blast and electric pig-iron, 401 
of depreciation, 299 
of desulphurizing, 307 

of electric pig-iron furnace installation, 239 
of the electrodes, 300 
installation of Girod furnace, 157 
of Heroult furnace, 141-142 
of Rochling-Rodenhauser furnace, 223 
of Stassano furnace, 120 
Creuzot, induction furnace of, 248 

Crucible furnace, heating with, costs compared to electric, 291 
Current density of electrodes (general), 82 
Girod furnace, 154 
Heroult furnace, 131 
Stassano furnace, 113 

Currents permitted in wires and cables, 29 
Cylinder winding, 165 

Davy's experiment, 3 

Delta connection, 63 

Deoxidation, 388-389 

Depreciation, 290 

Desulphurizing, cost of, 307 

Diagram of connections of a Kjellin furnace, 183 

Dinas, bricks, 281 

English or lime, bricks, 281 

German or clay, bricks, 281 
Direct current, applicability of, 68 
Dolomite, 282 

plant, 282 
Dynamo sheet-iron, 60 

408 INDEX 

Economical considerations, 257-277 
Economy of the electric shaft pig-iron furnace, 235 
Eddy currents, 59 

Efficiency, arc furnaces, influence of the electrode consumption on, 96-99, 132 
Efficiency of carbon electrodes, 84 
electric (general), 72 

Girod furnace, 152 

Heroult furnace, 136 

Kjellin furnace, 182, 187 

R. & R. furnace, 215 

Stassano furnace, 117 

electrode, according to C. A. Hansen and Carl Hering, 83-99 
graphite electrodes, 84 
shaft furnaces, electric, 236 
thermal, Girod furnace, 157 

Heroult furnace, 140 

R. & R. furnace, 223 

Stassano furnace, 120 
total, Girod furnace, 155 

Heroult furnace, 140 

Kjellin furnace, 191-192 

R. & R. furnace, 221-223 
Electric conditions of a Girod furnace, 147 

of a Heroult furnace, 128 

of a Kjellen furnace, 178-179 

of a Rochling-Rodenhauser furnace, 211 

of a Stassano furnace, 117 
furnace, demands of an ideal, 66-73 
furnaces, advantages of, 65-66, 74, 288-388 
pig-iron, characteristics of, 398 
power, cost of, 296 

steel, high quality characteristics of, 387 
steel production in Austria-Hungary, 293 
Electrode arrangement with Girod furnace, 145 

with Heroult furnace, 125 
cooling, 99 

consumed by Stassano furnace, 120-304 
consumption (general), 96-97 

with Girod furnace, 157 

with Heroult furnace, 129 

with shaft furnace, 238 

influence on furnace efficiency, 81-83 
cost, 300 
covering, 98 
cross section, comparison with Girod and Heroult furnaces, 154 

influence of, 80-82, 93, 129 
Electrode losses (general), 89 

lowest total, 86-90 

INDEX 409 

Electrode losses, with Girod furnace, 154 
with Heroult furnace, 129-130 
with the Stassano furnace, 114 
of pole plates R. R. furnace, 212 

pole plate consumption with Rochling-Rodenhauser furnaces, 211-212 
pole plates, with R. & R. furnace, 198 
regulation, general, 102 
Girod furnace, 151 
Heroult furnace, 125 
Stassano furnace, 116 
Electrodes, for arc furnaces. 80-102 
consuming, the, 89 
with Stassano furnace, 113-120 
Electro-metallurgy of iron, 319 
Electro-metals shaft furnace, 231, 332, 345 
Energy regulation of the Girod furnace, 152 
of the Heroult furnace, 125 
of the Kjellin furnace, 185 
of the R. & R. furnace, 214 
of the Stassano furnace, 113, 116, 118 
Expansion of the refractories, 280 

Ferranti, de, furnace, 165-166 
Ferro alloys, 376 
Ferro-chromium, 376 
Ferro-manganese, 376, 380 
Ferro-silicon, 376 

required for desulphurizing, 317 
Fluorspar, 377 

Flux, to lower melting point of refractories, 284 
Foucault currents, 59 
Frequency, 47, 48, 50 

with what, shall the electric furnace operate, 70 
Frick furnace, 169 

and Kjellin furnaces, differences, 170 
Furnace, at Allevard, 242-243 
refractories, 278-285 
size attainable with Girod type, 154 
with Heroult type, 140 
with Kjellin type, 190-191 
with R. & R. type, 221 
with Stassano type, 118 
system, its influence on the quality of steel made, 284 

Giffre furnace see Chapelet furnace, 242-243 
Gin, induction furnace of, 248 
resistance furnace of, 28 

arithmetical example of, 28-29 

410 INDEX 

Girod furnace, the, 144 

action of the heat, 150 

advantages of, 159 

applicability, 153 

arrangement, 145 

arrangement of electrodes, 145 

attainable size, 154 

circulation in the bath, 152 

comparison with an ideal furnace, 151 

cost of a furnace, 157 

crucible, 34 

current density in the electrodes, 154 

electrical conditions with, 147 

electrical efficiency, 152 

electrode cross section, 154 
losses, 154 

electrodes consumed, 157 

historical, 144 

influence of bottom electrodes, 153-149 

influence of the carbon electrodes. 155 

installations, 156, 159, 265 

kind of current used, 148 

licenses, giving, 159 

operation, 147 

power fluctuations, 151 

power used, 152-154 

refractories, 145 

regulating energy of, 152 

thermal efficiency, 157 

the tilting, 145-152 

total efficiency, 155 

operating cost, 305 

Graphite and carbon electrodes, comparison between, 90-98 
electrodes, efficiency of, 84 
heat conductivity and specific resistance, 90 
Gronwall arc furnace for steel, 249 
induction furnace for steel, 249 
Lindblad and Stalhane electric shaft xurnace, 231 

hearth furnace for smelting ore, 225, 231, 240 

Haanel and the Canadian Commission's report, 9 
Half schamotte bricks, 281 

silica bricks, 281 
Hearth arrangement, general, 73 

of the Girod furnace, 152 

of the Heroult furnace, 137 

of the Kjellin furnace, 188 

of the R. & R. furnace, 215 

INDEX 411 

Hearth arrangement, of the Stassano, 109, 118 
bottom with the Kjellin furnace, 175 
with the R. & R. furnace, 201 
with the Stassano furnace, no 
form and life of refractories, 285 
Heat action, 26-35 

conductivity of carbon, 91 

of graphite, 91 
losses, 86-87 

quantities, relations electrical and mechanical, 24-25 
required for ore reduction, 315 

Heating costs, comparison of blast and electric pig-iron furnace, 301-302 
of crucible and electric, 292 
of open hearth and electric, 290, 402 
influence present with arc furnaces, 38 

with electric furnaces in general, 73 
with induction furnaces, 172 
with the Girod furnace, 150 
with the Heroult furnace, 127 
with the Kjellin furnace, 183 
with the R. & R. furnace, 197 
with the Stassano furnace, 108, 115 
the Heroult furnace, 132 
the Kjellin furnace, 176-177 
the Rochling-Rodenhauser furnace, 206 
the Stassano furnace, 115 
Helberger, crucible furnace, 35 
Heraus, laboratory furnace, 34 
Heroult furnace, 124-125 

action of the heat, 127-128 

advantages of, 142-143 

applicability of, 138 

arc length of, 127 

arrangement of electrodes, 125 

attainable size, 140 

circulation in the bath, 137 

comparison with an ideal furnace, 133-134 

cost of a furnace, 141 

current density in electrodes, 129, 131 

current fluctuations, 126 

electric conditions, 128 

electrical efficiency, 136 

electrode cross section, 129, 131 

losses, 129, 130 

electrodes consumed, 129, 131, 140, 307 
historical, 121-123 
influence of carbon electrodes, 135 
installations, 124-261 

412 INDEX 

Heroult furnace, kind of current used, 141 
licenses, giving, 143 
operating cost of I5~ton, 138, 307, 308 
operation of, 132 
power fluctuations, 126 
power used, 133, 134 
regulating of energy, 125 
refractories and roof, 138, 307 
thermal efficiency, 140 
the tilting, 136 
total efficiency, 140 
Hiorth furnace, 250-251 
Historical in general, i-io 
of the Girod furnace, 144 
of the Heroult furnace, 12 1 
of the Kjellin furnace, 173 
of the Rochling-Rodenhauser furnace, 193 
of the Stassano furnace, 107 
Howe, criticism of, 286 
Hyteresis losses, 60 

Ideal electric furnace, compared to a Girod furnace, 151 
to a Heroult furnace, 133 
to a Kjellin furnace, 185 
to a R. & R. furnace, 212 
to a Stassano furnace, 116 
demands of, 65-72 
Impurities in the charge, getting rid of, 74 

in iron, 366 
Induced current, 161 

E.M.F. and its size, 163 
Induction, 49 

furnaces, combined, 172 

important points concerning, 171-172 
in general, 160-172 
of Gin, 248 

principal arrangement, 165 
pure, 165-169 
heating, 32 

characteristics, 164 
losses due to, phenomena, 51-53 
Installation at Aarau (Girod), 156 
at Allevard (Chapelet), 242-243 
at Bonn (Stassano), 113 
at Chicago (Heroult), 130 

at Dommeldingen (Rochling-Rodenhauser), 195 
at Essen (Frick), 170 
at Essen (Kjellin), 191 

INDEX 413 

Installation at Friedenshiitte (Nathusius), 246 

at Gysinge (Kjellin), 183 

at La Praz (Heroult), 123 

at Remscheid (Heroult), 124 

at Ugine (Girod), 159 

Volklingen (Rochling-Rodenhauser), 210-211 
Installations, statistics of electric steel furnace, 261-277 
pig-iron furnace, 241 
electric billet-heating furnaces, 256 
Instantaneous values, 48-50 
Iron, gray, avoidance bad heats in electric furnace, 381 

ore, 377 

reduction from iron pyrites, 326 

resistance of cold, 14 
of molten, 15 

Joule's law, 23 
Joule losses, 81 

Keller, arc furnace, 244 

pig-iron furnace, 226 
Kirchhoff's law, 19 
Kjellin furnace, 9, 169, 173, 192 

action of the heat, 183 

advantages of, 188 

applicability of, 188, 273, 274, 275 

attainable size, 190-191 

circulation in the bath, 188 

cooling of parts, 175, 176 

comparison with an ideal furnace, 185 

current fluctuations, 177, 185, 186 

electrical conditions, 178, 179 

electrical efficiency, 182, 187 

frequency, lowering of, 182 

historical, 173 

installations of, 272 

licenses, giving, for, 192 

operation of, 176, 184 

pinch effect, 190 

Poldihiitte, improved bottom, 190 

power factor, 178, 180, 181 
fluctuations, 186 
used, 192 

refractories and roof, 175 

regulation of energy, 185 

thermal efficiency, 192 

tilting type, 176, 187 

total efficiency, 191-192 

414 INDEX 

Kjellin furnace, transformer of, 173 

and Frick furnaces, the differences, 170 

Labor, 297 

Laboratory furnace of Borchers, 33 

of Heraus, 34 
Latent heat of fusion of pig iron, 323 

of slag, 323 

Laval, de, electric furnace of, 7 
Licenses, giving, for Girod furnaces, 159 

for Heroult furnaces, 143 

for Kjellin furnaces, 192 

for R. & R. furnaces, 224 

for Stassano furnaces, 120 
Lime, 376 

Dinas bricks, 281 
Line diagram, 47 
Lining, preventing attack, 386 
Loss, melting, 279 
Lyon furnace, 359 

Magnesite, 282 

bricks, 282 

Magnet, action between, and electric conductor, 40 
Magnets, action of two on each other, 40 
Magnetic lines of force, direction of, 41 

field of, cut by a conductor, 42 
of a coil, 43 

Magnetizing currents, 40 
Malleable iron castings, 381 
Manganese, absence of loss in electric furnace, 379 

ferro, less needed for deoxidation if liquid, 374 
Material charged, 287 

for furnace construction, 278 
Maximum values, 48 

Medal for Colby; his induction furnace, 166 
Melting pig iron, 379 
Mixer, electric furnace as a, 382 
Mortar, 282 
Motor effect, action of the electric current, 39 

Nathusius, arc furnace of, 245 
Neutral point, 63 

Ohm's Law, 11-12 
Ohm, the unit, 12 

Open hearth furnace and electric, comparison of their heating costs, 289,. 
290, 402 

INDEX 415 

Operating costs of the electric shaft and ordinary blast furnace, 302, 306 
of the Girod furnace, 152-154 
of the 15-ton Heroult's furnace, 307 
of the Rochling-Rodenhauser furnace, 308 
of the Stassano furnace, 304 

Operation, general requirements for electric furnace, 65-75 
of electric shaft furnaces, 233 
of the Girod furnace, 147 
of the Heroult furnace, 132 
of the Kjellin furnace, 176-184 
of the Rochling-Rodenhauser furnace, 221, 308 
of the Stassano furnace, 112-113 
Ore reduction, heat required for, 315 
smelting, 312-332 

criticism of, in the electric hearth furnace, 333 
in the electric shaft furnace, 335 

of Gronwall, Lindblad and Stalhane, 322 
in the hearth furnace of Gronwall, Lindblad and Stalhane, 345 

of R. & R., 323 

in the special furnace of Heroult, 339 
in the Stassano furnace, 319 
in the test furnace of Lyon, 359 
Oxygen, 373 

Parallel connection, 16-22 

Pepys' test, 4 

Period, periodicity, 47-50 

with what shall the electric furnace operate, 70 
Phase current, 63 
displacement, 54 

its influence in a. c. circuits, 55-58 
voltage, 63 
Phosphorus, 366 
Pichon, electric furnace of, 4 
Pig iron, carbon required per ton, 338 
Pinch effect, 44, 190 
Pipe casting, 380 
Pneumatic hammers, 285 

Poldihiitte, improvements in lining and bottom for Kjellin furnace, 190 
Pole plate electrodes with R. & R. furnace, 108, 300 
Power, apparent, 55 

cost of electric, per kw. hr., 296 
effective, 54 

factor, arithmetical example, 56 
influence of the, 178 

with Kjellin furnaces, influence of charge on, 178 
fluctuations, influence of, 71 

with the Girod furnace, 151 

416 INDEX 

Power fluctuations with the Heroult furnace, 126 

with the Kjellin furnace, 177 

with the RochlingrRodenhauser furnace, 113, 209 

with the Stassano furnace, 116 
generating cheap, 68, 258 
table for, 24-25 
three-phase circuit, 63-64 
used and its influences, 228 

with the electric pig-iron furnace, 239 

with the Girod furnace, 152-154 

with the Heroult furnace, 133-134 

with the Kjellin furnace, 192 

with the R. & R. furnace, 223 

with the Stassano furnace, 120 

Quality characteristics of electric iron and electric steel, 398 
of the steel, influence of the furnace type on, 286 
steel, making it in the electric furnace, 387 

Quartz, 280 

Quartzite, 280 

Quick melting, advantages of, 192 

Radiating furnaces, 79 

Reasons, economical, for introduction of electric furnace, 387 

Refining of pig iron, 383 

Refractories, cost of the, 287 

of the Girod furnace, 145 

of the Heroult furnace, 138, 307 

of the Kjellin furnace, 175 

of the R. & R. furnace, 201, 206, 218 

of the Stassano furnace, no, 115 
Refractory, durability, and hearth form, 284 

materials, 278-285 

material is called, when, 279 

mixtures, 282 

Regulating or auto-transformer, 214 
Resistance, apparent, 54 

of a conductor, 1 1 

of carbon, specific, 91 

change in graphite and carbon electrodes, 94-95 

of graphite, specific, 91 

heating, characteristics, 31-32 
direct and indirect, 25-32 

specific, 13-15 
Revolving furnace, 108-118 
Roasting furnace, 108, 118 

of ores, 350 
Rochling-Rodenhauser furnace, 197 

INDEX 417 

Rochling-Rodenhauser furnace, cost of heat, 196, 203 

advantages of, 193-194, 210 

applicability of, 219 

attainable size, 221 

circulation in the bath, 215 

comparison with ideal furnace, 212 

cooling of parts, 199-200, 212 

cost of a furnace, 223 

current fluctuation, 209 

electrical conditions in the, 211 

electrical efficiency, 214, 215 

historical, 193 

installations of, 274-6 

kind of current used, 223 

licenses, giving, for, 224 

operation of and cost, 308 

ore smelting in, 323 

power fluctuations, 113, 209 

power used, 223 

refractories and roof, 201, 206, 218 

regulation of energy, 43, 214 

scrap melting in, 207/208, 220, 223 

secondary circuit, 201, 205, 212 

shut down over Sunday, 209 

slag, absence of, in channels, 196, 207 

thermal efficiency, 223 

tilting type, 197, 215 

total efficiency, 221-223 

transformer of, 197 
Roof, life of, with arc and induction furnaces, 299 

with Girod furnace, 306 

with Heroult furnace, 138, 307 

with Rochling-Rodenhauser furnace, 299 

with Stassano furnace, 113 

Series connection, 18-24 

Shaft furnace, electric, 225-240 

economy of, 238-295 

efficiency, 238, 303, 304 

electrodes consumed, 235, 302, 304 

kind and quantity of carbon on, 239 

power consumption, 304 

of Gronwall, Lindblad and Stalhane, 231, 345 

of Heroult, 227, 338 

of Keller, 225-226 

of Lyon, 359 

of Stassano, 225 
Silicon, 372 

418 INDEX 

Silicon carbide bricks, 307 
Specific heat, 323 
Star connection, 62 
Stassano furnace, 106-120 

action of the heat, 114 

advantages of, 117 

applicability of, 118 

arc length, 114, 115 

arrangement of electrodes, iio-m 

attainable size, 118 

circulation in the bath, 109, 118 

comparison with an ideal furnace, 116 

cooling arrangements, in 

cost of a furnace, 120 

current density in electrodes, 113, 114, 118 

current fluctuations, 113, 116 

electric conditions, 117 

electrical efficiency, 117 

electrode cross section, 113 
losses, 114 

electrodes consumed, 120, 304 

energy regulation, 116 

hearth furnace, 117 

historical, 106 

installations, 113-267 

kind of current used, 117 

licenses given, 120 

operation of, 112, 113 

operation cost, 120 

ore smelting, 318 

power used, 120 

power fluctuations, 116 

regulating of energy, 113, 116, 118 

refractories and roof, no, 115 

rotating, 108 

shaft, or pig-iron furnace, 106 

thermal efficiency, 120 

tilting, 117 
Statistics of electric furnaces, 261-278 

pig-iron furnaces, 241 

heating furnaces, 255 
Steel, bad, non-expert handling, 396 
castings, electric, 118, 261-278 
Straying, 161, 171 

method of lessening the straying, 171 
'Sulphur, 368 

Tar, 282 

INDEX 419 

Taussig, electric furnace of, 8, 26 
Temperature coefficient, 14, 15 

regulation of electric furnaces, 72 

with electric heating, 73 

Thin-walled castings, gray iron, low phos., 380 
Three-phase current, 61 
Titanium, 375 
Tools used, 301 
Transformer (principal arrangement), 162-164 

iron, 60 

coefficient, 163 
Tube winding, 165 

Unburnt slag, 280 
Units, electrical, 12 

Values, instantaneous, 48-50 

Vanadium, 375 

Vector diagram, 51 

Very refractory, 279 

Volt, unit of electrical pressure, 12 

Water cooling, influence of, 75 
Watt component, 57 
Wattless component, 57 
Wattmeter, 57-64 
Work, table for delivered, 25 







London: CHAPMAN & HALL, Limited 
Montreal, Can.: RENOUF PUB. CO. 


Descriptive circulars sent on application. Books marked with an asterisk (*) are 
sold at net prices only. All books are bound in cloth unless otherwise stated. 


ARMSBY Principles of Animal Nutrition 8vo, $4 00 

BOWMAN Forest Physiography 8vo, *5 00 

BUDD and HANSEN American Horticultural Manual: 

Part I. Propagation, Culture, and Improvement 12mo, 1 50 

Part II. Systematic Pomology 12mo, 1 50 

ELLIOTT Engineering for Land Drainage 12mo, 2 00 

Practical Farm Drainage. (Second Edition, Rewritten) 12mo, 1 50 

FULLER Domestic Water Supplies for the Farm 8vo, *1 50 

GRAVES Forest Mensuration 8vo, 4 00 

Principles of Handling Woodlands Small 8vo, *1 50 

GREEN Principles of American Forestry 12mo, 1 50 

GROTENFELT Principles of Modern Dairy Practice. (WoLL.) 12mo, 2 00 

HAWLEY and HAWES Forestry in New England 8vo, *3 50 

HERRICK Denatured or Industrial Alcohol 8vo, *4 00 

KEMP and WAUGH Landscape Gardening. (New Edition, Rewritten). . 12mo, *1 50 

LARSEN and WHITE Dairy Technology Small 8vo, *1 50 

McKAY and LARSEN Principles and Practice of Butter-making 8vo, *1 50 

MAYNARD Landscape Gardening as Applied to Home Decoration . . . 12mo, 1 50 
RECORD Identification of the Economic Woods of the United States.. .8vo, *1 25 
RECKNAGEL Theory and Practice of Working Plans (Forest Organi- 
zation) (In Press.) 

SANDERSON Insect Pests of Farm, Garden, and Orchard Small 8vo, *3 00 

Insects Injurious to Staple Crops 12mo, 1 50 

SCHWARZ Longleaf Pine in Virgin Forest 12mo, *1 25 

SOLOTAROFF -Field Book for Street-tree Mapping 12mo, *0 75 

In lots of one dozen % . *8 00 

. Shade Trees in Towns and Cities .' 8vo, *3 00 

STOCKBRIDGE Rocks and Soils 8vo, 2 50 

WINTON Microscopy of Vegetable Foods Large 8vo, 7 50 

WOLL Handbook for Farmer i and Dairymen IGmo, 1 50 


ATKINSON Orientation of Buildings or Planning for Sunlight 8vo, *2 00 

BALDWIN Steam Heating for Buildings 12mo, 2 50 

BERG Buildings and Structures of American Railroads 4to, 5 00 


BIRKMIRE Architectural Iron and Steel 8vo, $3 50 

Compound Riveted Girders as Applied in Buildings 8vo, 2 00 

Planning and Construction of High Office Buildings 8vo, 3 50 

Skeleton Construction in Buildings 8vo, 3 00 

BRIGGS Modern American School Buildings 8vo, 4 00 

BYRNE Inspection of Materials and Workmanship Employed in Construc- 
tion IGmo, 3 00 

CARPENTER Heating and Ventilating of Buildings 8vo, 4 00 

CORTHELL Allowable Pressure on Deep Foundations 12mo, *1 25 

ECKEL Building Stones and Clays 8vo, *3 00 

FREITAG Architectural Engineering 8vo, 3 50 

Fire Prevention and Fire Protection IGmo, mor., *4 00 

Fireproofing of Steel Buildings 8vo, 2 50 

GERHARD Guide to Sanitary Inspections 12mo, 1 50 

Modern Baths and Bath Houses 8vo, *3 00 

Sanitation of Public Buildings 12mo, 1 50 

Theatre Fires and Panics 12mo, 1 50 

The Water Supply, Sewerage and Plumbing of Modern City Buildings, 

8vo, *4 00 

GREENE Elements of Heating and Ventilation 8vo, *2 50 

JOHNSON Statics by Algebraic and Graphic Methods 8vo, 2 00 

KELLAWAY How to Lay Out Suburban Home Grounds 8vo, 2 00 

KIDDER Architects' and Builders' Pocket-book 16mo, mor., 5 00 

MERRILL Stones for Building and Decoration 8vo, 5 00 

MONCKTON Stair-building 4to, 4 00 

PATTON Practical Treatise on Foundations 8vo, 5 00 

PEABODY Naval Architecture 8vo, 7 50 

RICE Concrete-block Manufacture 8vo, 2 00 

RICHEY Handbook for Superintendents of Construction 16mo, mor., 4 00 

Building Foreman's Pocket Book and Ready Reference.. . .IGmo, mor., 5 00 
Building Mechanics' Ready Reference Series: * 

Carpenters' and Woodworkers' Edition IGmo, mor., *1 50 

Cement Workers' and Plasterers' Edition IGmo, mor., *1 50 

Plumbers', Steam-Fitters', and Tinners' Edition IGmo, mor., *1 50 

Stone- and Brick-masons' Edition IGmo, mor., *1 50 

RIES Building Stones and Clay Products 8vo, *3 00 

SABIN House Painting (Glazing, Paper Hanging and Whitewashing) . 12mo, 1 00 

SIEBERT and BIGGIN Modern Stone-cutting and Masonry 8vo, 1 50 

SNOW Principal Species of Wood 8vo, 3 50 

WAIT Engineering and Architectural Jurisprudence 8vo, 6 00 

Sheep, 6 50 

Law of Contracts 8vo, 3 00 

Law of Operations Preliminary to Construction in Engineering and 

Architecture 8vo, 5 00 

Sheep, 5 50 

WILSON Air Conditioning 12mo, 1 50 

WORCESTER and ATKINSON Small Hospitals, Establishment and Mainte- 
nance, Suggestions for Hospital Architecture, with Plans for a Small 
Hospital 12mo, 1 25 


BETTS Lead Refining by Electrolysis 8vo, 4 00 

FLETCHER Practical Instructions in Quantitative Assaying with the Blow- 


IGmo, mor., 1 50 

FURMAN and PARDOE Manual of Practical Assaying 8vo, 3 00 

LODGE Notes on Assaying and Metallurgical Laboratory Experiments. 8vo, 3 00 

Low Technical Methods of Ore Analysis 8vo, 3 00 

MILLER Cyanide Process 12mo, 1 00 

Manual of Assaying 12mo, 1 00 

MINET Production of Aluminum and its Industrial Use. (WALDO). . 12mo, 2 50 
PRICE and MEADE The Technical Analysis of Brass and the Non-Ferrous 

Alloys 12mo, *2 00 

RICKETTS and MILLER Notes on Assaying 8vo, 3 00 

ROBINE and LENGLEN Cyanide Industry. (LE CLERC.) 8vo, 4 00 

SEAMON Manual for Assayers and Chemists , , .Small 8vo, *2 50 


ULKE Modern Electrolytic Copper Refining 8vo, $3 00 

WILSON Chlorination Process 12mo, 1 50 

Cyanide Processes 12mo, 1 50 


COMSTOCK Field Astronomy for Engineers 8vo, 2 50 

CRAIG Azimuth 4to, 3 50 

CRANDALL Text-book on Geodesy and Least Squares 8vo, 3 00 

DOOLITTLE Treatise on Practical Astronomy Svo, 4 00 

HAYFORD Text-book of Geodetic Astronomy Svo, 3 00 

HOSMER Azimuth IGmo, mor., 1 00 

Text-book on Practical Astronomy Svo, *2 00 

MERRIMAN Elements of Precise Surveying and Geodesy Svo, 2 50 

MICHIE and HARLOW Practical Astronomy Svo, *3 00 

RUST Ex-meridian Altitude, Azimuth and Star-Finding Tables Svo, 5 00 

WHITE Elements of Theoretical and Descriptive Astronomy 12mo, *2 00 


COHNHEIM Enzymes 12mo, *1 50 

DAVENPORT Statistical Methods with Special Reference to Biological 

Variation 16mo, mor., 1 50 

EFFRONT and PRESCOTT Enzymes and Their Applications Svo, 3 00 

EULER and POPE General Chemistry of the Enzymes Svo, *3 00 

MAST Light and Behavior of Organisms Svo, *2 50' 

PRESCOTT and WINSLOW Elements of Water Bacteriology, with Special 

Reference to Sanitary Water Analysis 12mo, 1 50 

WARD and WHIPPLE Freshwater Biology (In Press.) 

WHIPPLE The Microscopy of Drinking Water Svo, 3 50 

WINSLOW The Systematic Relationship of the Coccaceae Small Svo, 2 50- 


ABDERHALDEN Physiological Chemistry in Thirty Lectures. (HALL and 

DEFREN) Svo, *5 00" 

ABEGG Theory of Electrolytic Dissociation. (VON ENDE.) 12mo, *1 25 

ALEXEYEFF General Principles of Organic Syntheses. (MATTHEWS.). . Svo, 3 00 
ALLEN Exercises in General Chemistry, Loose Leaf Laboratory Manual. 

Oblong 4to, paper, *1 00 

Tables for Iron Analysis Svo, 3 00 

'ARMSBY Principles of Animal Nutrition Svo, 4 00 

ARNOLD Compendium of Chemistry. (MANDEL.) Small Svo, 3 50 


Hartford Meeting, 1906 Svo, 3 00 

Jamestown Meeting, 1907 Svo, 3 00 

AUSTEN Notes for Chemical Students 12mo, 1 50 

BERNADOU Smokeless Powder. Nitro-cellulose, and Theory of the Cellu- 
lose Molecule 12mo, 2 50 

BILTZ Introduction to Inorganic Chemistry. (HALL atid PHELAN.).. 12mo, *1 25 
Laboratory Methods of Inorganic Chemistry. (HALL and BLANCHARD.) 

Svo, 3 00 
BINGHAM and WHITE Laboratory Manual of Inorganic Chemistry. . . . 12mo, *1 00 

BLANCHARD Synthetic Inorganic Chemistry , . . . 12mo, *1 00 

BOTTLER German and American Varnish Making. (SABIN.).. . Small Svo, *3 50 

BROWNE Handbook of Sugar Analysis Svo, *6 00 

Sugar Tables for Laboratory Use Svo, *1 25 

BROWNING Introduction to the Rarer Elements Svo, *1 50 

BRUNSWIG Explosives. (MUNROE and KIBLER) Small Svo, *3 00 

CLAASSEN Beet-sugar Manufacture. (HALL and ROLFE.) Svo, *3 Oft 

CLASSEN Quantitative Analysis by Electrolysis. (HALL.) (In Press.) 


COHN Indicators and Test-papers 12mo, $2 00 

Tests and Reagents -. Svo, 3 00 

COOPER Constitutional Analysis by Physico-chemical Methods (Inorganic), 

(In Press.) 

DANNEEL Electrochemistry. (MERRIAM.) 12mo, *1 25 

DANNERTH Methods of Textile Chemistry 12mo, 2 00 

DUHEM Thermodynamics and Chemistry. (BURGESS.) 8vo, 4 00 

EISSLER Modern High Explosives 8vo, 4 00 

EKELEY Laboratory Manual of Inorganic Chemistry 12mo, *1 00 

FLETCHER Practical Instructions in Quantitative Assaying with the Blow- 
pipe * 16mo, mor., 1 50 

FOWLER Sewage Works Analyses 12mo, 2 00 

FRESENIUS Manual of Qualitative Chemical Analysis. (WELLS.) 8vo, 5 00 

Manual of Qualitative Chemical Analysis. Part I. Descriptive. 

(WELLS.) 8vo, 3 00 

Quantitative Chemical Analysis. (ConN.) 2 vols 8vo, 12 50 

When Sold Separately, Vol. I, $6. Vol. II, $8. 

FUERTES Water and Public Health 12mo, 1 50 

FULLER Qualitative Analysis of Medicinal Preparations 12mo, *1 50 

FURMAN and PARDOE Manual of Practical Assaying 8vo, 3 00 

GETMAN Exercises in Physical Chemistry 12mo, *2 00 

GILL Gas and Fuel Analysis for Engineers 12mo, 1 25 

GOOCH Methods in Chemical Analysis 8vo, *4 00 

and BROWNING Outlines of Qualitative Chemical Analysis. . Small 8vo. *1 25 

GROTENFELT Principles of Modern Dairy Practice. (WoLL.) 12mo, 2 00 

GROTH Introduction to Chemical Crystallography. (MARSHALL.). .. 12mo, 1 25 
HAMMARSTEN Text-book of Physiological Chemistry. (MANDEL.). . . .8vo, *4 00 

HANAUSEK Microscopy of Technical Products. (WINTON.) 8vo, 5 00 

HASKINS Organic Chemistry 12mo, *2 00 

HERRICK Denatured or Industrial Alcohol 8vo, *4 00 

HINDS Inorganic Chemistry 8vo, 3 00 

Laboratory Manual for Students 12mo, *1 00 

HOLLEMAN Laboratory Manual of Organic Chemistry for Beginners. 

(WALKER.) 12mo, *1 00 

Text-book of Inorganic Chemistry. (COOPER.) 8vo, 2 50 

Text-book of Organic Chemistry. (WALKER AND MOTT.) 8vo, 2 50 

(EKELEY) Laboratory Manual to Accompany Holleman's Text-book of 

Inorganic Chemistry 12mo, *1 00 

HOLLEY Analysis of Paint and Varnish Products. . . . . Small, 8vo, *2 50 

Lead and Zinc Pigments Small 8vo, *3 00 

HOPKINS Oil-chemists' Handbook 8vo, 3 00 

JACKSON Directions for Laboratory Work in Physiological Chemistry. .8vo, 1 25 
JOHNSON Rapid Methods for the Chemical Analysis of Special Steels, Steel- 
making Alloys and Graphite Small 8vo, 3 00 

LANDAUER Spectrum Analysis. (TINGLE.) 8vo, 3 00 

LASSAR-COHN Application of Some General Reactions to Investigations in 

Organic Chemistry. (TINGLE.) 12mo, 1 00 

LEACH Food Inspection and Analysis Svo, 7 50 

LOB Electrochemistry of Organic Compounds. (LoRENZ.) Svo, 3 00 

LODGE Notes on Assaying and Metallurgical Laboratory Experiments . . Svo, 3 00 

Low Technical Method of Ore Analysis Svo, 3 00 

LOWE Paint for Steel Structures 12mo, 1 00 

LUNGE Techno-chemical Analysis. (COHN.) 12mo, 1 00 

McKAY and LARSEN Principles and Practice of Butter-making Svo, *1 50 

MAIRE Modern Pigments and their Vehicles 12mo, 2 00 

MANDEL Handbook for Bio-chemical Laboratory 12mo, 1 50 

MARTIN Laboratory Guide to Qualitative Analysis with the Blowpipe. 12mo, *0 60 
MASON Examination of Water. (Chemical and Bacteriological.).. .. 12mo, 1 25 

Water-supply. (Considered Principally from a Sanitary Standpoint.) 

Svo, 4 00 

MATHEWSON First Principles of Chemical Theory Svo, *1 00 

MATTHEWS Laboratory Manual of Dyeing and Textile Chemistry Svo, 3 50 

Textile Fibres Svo, 4 00 

MEYER Determination of Radicles in Carbon Compounds. (TINGLE.) 

Third Edition 12mo, *1 25 

MILLER Cyanide Process 12mo, 1 00 

Manual of Assaying 12mo, 1 00 


MINET Production of Aluminum and its Industrial Use. (WALDO.) . . 12mo, $2 50 
MITTELSTA.EDT Technical Calculations for Sugar Works. (BOURBAKIS.) 

12mo, *1 50 

MIXTER Elementary Text-book of Chemistry 12mo, 1 50 

MORGAN Elements of Physical Chemistry 12mo, 3 00 

Physical Chemistry for Electrical Engineers 12mo, *1 50 

MOORE Experiments in Organic Chemistry 12mo, *0 50 

Outlines of Organic Chemistry 12mo, *1 50 

MORSE Calculations used in Cane-sugar Factories IGmo, mor., 1 50 

MUIR History of Chemical Theories and Laws 8vo, *4 00 

MULLIKEN General Method for the Identification of Pure Organic Com- <\ 

pounds. Vol. I. Compounds of Carbon with Hydrogen and 
Oxygen Large 8vo, 5 00 

Vol. II. Nitrogenous Compounds (In Preparation.) 

Vol. III. The Commercial Dyestuffs Large 8vo, 5 00 

NELSON Analysis of Drugs and Medicines 12mo, *3 00 

OSTWALD Conversations on Chemistry. Part One. (RAMSEY.). ... 12mo, 1 50 

Part Two. (TuRNBULL.).12mo, 2 00 

Introduction to Chemistry. (HALL and WILLIAMS.) Small 8vo, *1 50 

OWEN and STANDAGE Dyeing and Cleaning of Textile Fabrics 12mo, 2 00 

PALMER Practical Test Book of Chemistry 12mo, *1 00 

PAULI Physical Chemistry in the Service of Medicine. (FISCHER.).. 12mo, *1 25 

PICTET Alkaloids and their Chemical Constitution. (BiDDLE.) 8vo, 5 00 

PRESCOTT and WINSLOW Elements of Water Bacteriology, with Special 

Reference to Sanitary Water Analysis 12mo, 1 50 

REISIG Guide to Piece-Dyeing 8vo, *25 00 

RICHARDS and WOODMAN Air, Water, and Food from a Sanitary Stand- 
point 8vo, 2 00 

RICKETTS and MILLER Notes on Assaying 8vo, 3 00 

RIDEAL Disinfection and the Preservation of Food 8vo, 4 00 

RIGGS Elementary Manual for the Chemical Laboratory 8vo, 1 25 

ROBINE and LENGLEN Cyanide Industry. (LE CLERC.) 8vo, 4 00 

RUDDIMAN Incompatibilities in Prescriptions 8vo, 2 00 

Whys in Pharmacy 12mo, 1 00 

RUER Elements of Metallography. (MATHEWSON.) 8vo, *3 00 

SABIN Industrial and Artistic Technology of Paint and Varnish 8vo, 3 00 

SALKOWSKI Physiological and Pathological Chemistry. (Orndorff.). . .8vo, 2 50 
SCHIMPF Essentials of Volumetric Analysis Small 8vo, *1 50 

Manual of Volumetric Analysis 8vo, 5 00 

Qualitative Chemical Analysis 8vo, *1 25 

SEAMON Manual for Assayers and Chemists Small 8vo, *2 50 

SMITH Lecture Notes on Chemistry for Dental Students 8vo, *2 50 

SPENCER Handbook for Cane Sugar Manufacturers IGmo, mor., 3 00 

Handbook for Chemists of Beet-sugar Houses IGmo, mor., 3 00 

STOCKBRIDGE Rocks and Soils 8vo, 2 50 

STONE Practical Testing of Gas and Gas Meters 8vo, 3 50 

TILLMAN Descriptive General Chemistry 8vo, *3 00 

Elementary Lessons in Heat 8vo, *1 50 

TREADWELL Qualitative Analysis. (HALL.) 8vo, 3 00 

Quantitative Analysis. (HALL.) 8vo, 4 00 

TURNEAURE and RUSSELL Public Water-supplies 8vo, 5 00 

VAN DEVENTER Physical Chemistry for Beginners. (BOLTWOOD.) . . 12mo, 1 50 
VENABLE Methods and Devices for Bacterial Treatment of Sewage. . .8vo, 3 00 

WARD and WHIFFLE Freshwater Biology (In Press.) 

WARE Beet-sugar Manufacture and Refining. Vol. I 8vo, 4 00 

Vol. II 8vo, 5 00 

WASHINGTON Manual of the Chemical Analysis of Rocks 8vo, 2 00 

WEAVER Military Explosives _ 8vo, *3 00 

WELLS Laboratory Guide in Qualitative Chemical Analysis 8vo, 1 50 

Short Course in Inorganic Qualitative Chemical Analysis for Engineering 

Students 12mo, 1 50 

Text-book of Chemical Arithmetic 12mo, 1 25 

WHIFFLE Microscopy of Drinking-water 8vo, 3 50 

WILSON Chlorination Process - 12mo, 1 50 

Cyanide Processes 12mo, 1 50 

WINTON Microscopy of Vegetable Foods 8vo, 7 50 

ZSIGMONDY Colloids and the Ultramicroscope. (ALEXANDER.) . Large 12mo, 3 00 





Editor-in-chief.) 16mo, mor.,*S5 00 

BAKER Engineers' Surveying Instruments 12mo, 3 00 

BIXBY Graphical Computing Table Paper 19j X24| inches, 25 

BREED and HOSMER Principles and Practice of Surveying. 

Vol. I. Elementary Surveying 8vo, 3 00. 

Vol. II. Higher Surveying 8vo, 2 50 

BURR Ancient and Modern Engineering and the Isthmian Canal 8vo, *3 50 

COMSTOCK Field Astronomy for Engineers 8vo, 2 50 

CORTHELL Allowable Pressure on Deep Foundations 12mo, *1 25 

CRANDALL Text-book on Geodesy and Least Squares 8vo, 3 00 

DAVIS Elevation and Stadia Tables 8vo, 1 00 

ELLIOTT Engineering for Land Drainage 12mo, 2 00 

FIEBEGER Treatise on Civil Engineering Svo, *5 00 

FLEMER Phototopographic Methods and Instruments Svo, 5 00 

FOLWELL Sewerage. (Designing and Maintenance.) Svo, 3 00 

FREITAG Architectural Engineering Svo, 3 50 

HAUCH and RICE Tables of Quantities for Preliminary Estimates. . . 12mo, *1 25 

HAYFORD Text-book of Geodetic Astronomy Svo, 3 00 

HERING Ready Reference Tables (Conversion Factors.) 16mo, mor., 2 50 

HOSMER Azimuth 16mo, mor., 1 00 

Text-book on Practical Astronomy .' . . .Svo, *2 00 

HOWE Retaining Walls for Earth 12mo, 1 25 

IVES Adjustments of the Engineer's Transit and Level IGmo, bds., *0 25 

IVES and HILTS Problems in Surveying, Railroad Surveying and Geod- 
esy 16mo, mor., 1 50 

JOHNSON (J. B.) and SMITH Theory and Practice of Surveying . Small Svo, *3 50 

JOHNSON (L. J.) Statics by Algebraic and Graphic Methods Svo, 2 00 

KINNICUTT, WINSLOW and PRATT Sewage Disposal Svo, *3 00 

MAHAN Descriptive Geometry Svo, *1 50 

MERRIMAN Elements of Precise Surveying and Geodesy Svo, 2 50 

MERRIMAN and BROOKS Handbook for Surveyors 16mo, mor., 2 00 

NUGENT Plane Surveying Svo, 3 50 

OGDEN Sewer Construction Svo, 3 00 

Sewer Design 12mo, 2 00 

OGDEN and CLEVELAND Practical Methods of Sewage Disposal for Resi- 
dences, Hotels, and Institutions .' Svo, *1 50 

PARSONS Disposal of Municipal Refuse Svo, 2 00 

PATTON Treatise on Civil Engineering Svo, half leather, 7 50 

REED Topographical Drawing and Sketching 4to, 5 00 

RIEMER Shaft-sinking under Difficult Conditions. ( CORNING and PEELE.) 

Svo, 3 00 

SIEBERT and BIGGIN Modern Stone-cutting and Masonry Svo, 1 50 

SMITH Manual of Topographical Drawing. (McMiLLAN.) Svo, 2 50 

SOPER Air and Ventilation of Subways 12mo, 2 50 

TRACY Exercises in Surveying 12mo, mor., *1 00 

Plane Surveying 16mo, mor., 3 00 

VENABLE Garbage Crematories in America Svo, 2 00 

Methods and Devices for Bacterial Treatment of Sewage Svo, 3 00 

WAIT Engineering and Architectural Jurisprudence Svo, 6 00 

Sheep, 6 50 

Law of Contracts Svo, 3 00 

Law of Operations Preliminary to Construction in Engineering and 

Architecture Svo, 5 00 

Sheep, 5 50 

WARREN Stereotomy Problems in Stone-cutting Svo, 2 50 

WATERBURY Vest-Pocket Hand-book of Mathematics for Engineers. 

21 X5f inches, mor., *1 00 

Enlarged Edition, Including Tables mor., *1 50 

WEBB Problems in the Use and Adjustment of Engineering Instruments. 

16mo, mor., 1 25 
WILSON Topographic, Trigonometric and Geodetic Surveying Svo, 3 50 


BISHOP Structural Details of Hip and Valley Rafters... Oblong large 8vo. *$1 75 
BOLLER Practical Treatise on the Construction of Iron Highway Bridges 

8vo, 2 00 

Thames River Bridge Oblong paper, *5 00 

BURR and FALK Design and Construction of Metallic Bridges 8vo, 5 00 

Influence Lines for Bridge and Roof Computations 8vo, 3 00 

Du Bois Mechanics of Engineering. Vol. II Small 4to, 10 00 

FOSTER Treatise on Wooden Trestle Bridges 4to, 5 00 

FOWLER Ordinary Foundations 8vo, 3 50 

GREENE Arches in Wood, Iron, and Stone 8vo, 2 50 

Bridge Trusses 8vo, 2 50 

Roof Trusses 8vo, 1 25 

GRIMM Secondary Stresses in Bridge Trusses 8vo, 2 50 

HELLER Stresses in Structures and the Accompanying Deformations. .8vo, 3 00 

HOWE Design of Simple Roof -trusses in Wood and Steel 8vo, 2 00 

Symmetrical Masonry Arches 8vo, 2 50 

Treatise on Arches 8vo, 4 00 

HUDSON Deflections and Statically Indeterminate Stresses Small 4to, *3 50 

Plate Girder Design 8vo, *1 50 

JACOBY Structural Details, or Elements of Design in Heavy Framing, 8vo, *2 25 
JOHNSON, BRYAN and TURNEAURE Theory and Practice in the Designing of 
Modern Framed Structures. New Edition. 

Part I. Stresses in Simple Structures 8vo, *3 00 

Part II. Statically Indeterminate Structures and Secondary Stresses 

8vo, *4 00 
MERRIMAN and JACOBY Text-book on Roofs and Bridges: 

Part I. Stresses in Simple Trusses 8vo, 2 50 

Part II. Graphic Statics 8vo, 2 50 

Part III. Bridge Design 8vo, 2 50 

Part IV. Higher Structures 8vo, 2 50 

RICKER Design and Construction of Roofs 8vo, 5 00 

SONDERICKER Graphic Statics, with Applications to Trusses, Beams, and 

Arches 8vo, *2 00 

WADDELL De Pontibu?, Pocket-book for Bridge Engineers. . . . IGmo, mor., 2 00 
Specifications for Steel Bridges 12mo, *0 50 


BARNES Ice Formation 8vo, 3 00 

BAZIN Experiments upon the Contraction of the Liquid Vein Issuing from 

an Orifice. (TRAUTWINE.) 8vo, 2 00 

BO.VEY Treatise on Hydraulics 8vo, 5 00 

CHURCH Diagrams of Mean Velocity of Water in Open Channels. 

Oblong 4to, paper, 1 50 

Hydraulic Motors 8vo, 2 00 

Mechanics of Fluids (Being Part IV of Mechanics of Engineering) . .8vo, 3 00 

COFFIN Graphical Solution of Hydraulic Problems IGmo, mor., 2 50 

FLATHER Dynamometers, and the Measurement of Power 12mo, 3 00 

FOLWELL Water-supply Engineering 8vo, 4 00 

FRIZELL Water-power 8vo, 5 00 

FUERTES Water and Public Health 12mo, 1 50 

FULLER Domestic Water Supplies for the Farm 8vo, *1 50 

GANGUILLET and KUTTER General Formula for the Uniform Flow of Water 

in Rivers and Other Channels. (HERING and TRAUTWINE.).. .8vo, 4 00 

HAZEN Clean Water and How to Get It . % Small 8vo, 1 50 

Filtration of Public Water-supplies ' 8vo, 3 00 

HAZELHURST Towers and Tanks for Water-works 8vo, 2 50 

HERSCHEL 115 Experiments on the Carrying Capacity of Large, Riveted, 

Metal Conduits 8vo, 2 00 

HOYT and GROVER River Discharge 8vo, 2 00 

HUBBARD and KIERSTED Water-works Management and Maintenance, 8vo, 4 00 
LYNDON Development and Electrical Distribution of Water Power. .8vo, *3 00 
MASON Water-supply. (Considered Principally from a Sanitary Stand- 
point.) 8vo, 4 00 

MERRIMAN Elements of Hydraulics 12mo, *1 00 

Treatise on Hydraulics. 9th Edition, Rewritten 8vo, *4 00 


MOLITOR Hydraulics of Rivers, Weirs and Sluices 8vo,*S2 00 

MORRISON and BRODIE High Masonry Dam Design 8vo,' *1 50 

SCHUYLER Reservoirs for Irrigation, Water-power, and Domestic Water 

supply. Second Edition, Revised and Enlarged Large 8vo, 6 00 

THOMAS and WATT Improvement of Rivers 4to. *6 00 

TURNEAURE and RUSSELL Public Water-supplies 8vo, 5 00 

WEGMANN Design and Construction of Dams. 6th Ed., enlarged 4to, *6 00 

Water Supply of the City of New York from 1658 to 1895 4to, 10 00 

WHIPPLE Value of Pure Water Small 8vo, 1 00 

WILLIAMS and HAZEN Hydraulic Tables 8vo, 1 50 

WILSON Irrigation Engineering 8vo, 4 00 

WOOD Turbines 8vo, 2 50 


BAKER Roads and Pavements 8vo, 5 00 

Treatise on Masonry Construction 8vo, 5 00 

BLACK United States Public Works Oblong 4to, 5 00 

BLANCHARD Bituminous Surfaces and Bituminous Pavements. (In Preparation.) 

and DROWNE Highway Engineering, as Presented at the Second 

International Road Congress, Brussels, 1910 8vo, *2 00 

Text-book on Highway Engineering (In Press.) 

BOTTLER German and American Varnish Making. (SABIN.) . . . Small 8vo, *3 50 

BURR Elasticity and Resistance of the Materials of Engineering 8vo, 7 50 

BYRNE Highway Construction 8vo, 5 00 

Inspection of the Materials and Workmanship Employed in Construction. 

16mo, 3 00 
CHURCH Mechanics of Engineering Svo, 6 00 

Mechanics of Solids (Being Parts I, II, III of Mechanics of Engineer- 
ing) Svo, 4 50 

Mechanics of Fluids (Being Part IV of Mechanics of Engineering) . Svo, 3 00 
Du Bois Mechanics of Engineering: | 

Vol. I. Kinematics, Statics, Kinetics Small 4to, 7^50 

Vol. II. The Stresses in Framed Structures, Strength of Materials and 

Theory of Flexures Small 4to, 10 00 

ECKEL Building Stones and Clays ? Svo, *3 00 

Cements, Limes, and Plasters Svo, *6 00 

FOWLER Ordinary Foundations Svo, 3 50 

FULLER and JOHNSTON Applied Mechanics: 

Vol. I. Theory of Statics and Kinetics (In Press.) 

Vol. II. Strength of Materials (In Preparation.) 

GREENE Structural Mechanics ; Svo, *2 50 

HOLLEY Analysis of Paint and Varnish Products Small Svo, *2 50 

Lead and Zinc Pigments Small Svo, *3 '00 

HUBBARD Dust Preventives and Road Binders Svo, *3 00 

JOHNSON (J. B.) Materials of Construction Large Svo, 6 00 

KEEP Cast Iron Svo, 2 50 

KING Elements of the Mechanics of Materials and of Power of Transmis- 
sion Svo, *2 50 

LANZA Applied Mechanics Svo, 7 50 

LOWE Paints for Steel Structures 12mo, 1 00 

MAIRE Modern Pigments and their Vehicles 12mo, 2 00 

M A.URER Technical Mechanics Svo, 4 00 

MERRILL Stones for Building and Decoration Svo, 5 00 

MERRIMAN Mechanics of Materials Svo, 5 00 

Strength of Materials 12mo, *1 00 

METCALF Steel. A Manual for Steel-users 12mo, 2 00 

MORRISON Highway Engineering Svo, 2 50 

MURDOCK Strength of Materials 12mo, *2 00 

PATTON Practical Treatise on Foundations Svo, 5 00 

RICE Concrete Block Manufacture Svo, 2 00 

RICHARDSON Modern Asphalt Pavement Svo, 3 00 

RICHEY Building Foreman's Pocket Book and Ready Reference . 16mo,mor. , 5 00 

Cement Workers' and Plasterers' Edition (Building Mechanics' Ready 

Reference Series) 16mo, mor., *1 50 

Handbook for Superintendents of Construction 16mo, mor., 4 00 

Stone and Brick Masons' Edition (Building Mechanics' Ready Reference 

Series) 16mo. mor., *1 50 


RIES Building Stones and Clay Products 8vo.*$3 00 

Clays: Their Occurrence, Properties, and Uses 8vo, *5 00 

and LEIGHTON History of the Clay-working Industry of the United 

States 8vo, *2 50 

SABIN Industrial and Artistic Technology of Paint and Varnish 8vo, 3 00 

SMITH Strength of Material . 12mo, *1 25 

SNOW Principal Species of Wood 8vo, 3 50 

SPALDING Hydraulic Cement 12mo, 2 00 

Text-book on Road and Pavements 12mo, *2 00 

TAYLOR and THOMPSON Concrete Costs Small 8vo, *5 00 

Extracts on Reinforced Concrete Design 8vo, *2 00 

Treatise on Concrete, Plain and Reinforced 8vo, 5 00 

THURSTON Materials of Engineering. In Three Parts 8vo, 8 00 

Part I. Non-metallic Materials of Engineering and Metallurgy. . .8vo, 2 00 

Part II. Iron and Steel 8vo, 3 50 

Part III. A Treatise on Brasses, Bronzes, and Other Alloys and their 

Constituents 8vo, 2 50 

TILLSON Street Pavements and Paving Materials 8vo, *4 00 

TURNEAURE and MAURER Principles of Reinforced Concrete Construction. 

8vo, 3 50 

WATERBURY Cement Laboratory Manual 12mo, 1 00 

Laboratory Manual for Testing Materials of Construction 12mo, *1 50 

WOOD (DE V.) Treatise on the Resistance of Materials, and an Appendix on 

the Preservation of Timber Svo, 2 00 

(M. P.) Rustless Coatings: Corrosion and Electrolysis of Iron and 

Steel Svo, 4 00 


BERG Buildings and Structures of American Railroads 4to, 5 00 

BROOKS Handbook of Street Railroad Location -. 16mo, mor., 1 50 

BURT Railway Station Service 12mo, *2 00 

BUTTS Civil Engineer's Field-book 16mo, mor., 2 50 

CRANDALL Railway and Other Earthwork Tables Svo, 1 50 

and'BARNES Railroad Surveying 16mo, mor., 2 00 

CROCKETT Methods for Earthwork Computations Svo, *1 50 

DREDGE History of the Pennsylvania Railroad. (1879) Paper, 5 00 

FISH Earthwork Haul and Overhaul (In Press.) 

FISHER Table of Cubic Yards Cardboard, 25 

GILBERT, WIGHTMAN and SAUNDERS Subways and Tunnels of New York. 

Svo, *4 00 

GODWIN Railroad Engineers' Field-book and Explorers' Guide. .16mo, mor., 2 50 
HUDSON Tables for Calculating the Cubic Contents of Excavations and 

Embankments Svo, 1 00 

IVES and HILTS Problems in Surveying, Railroad Surveying and Geodesy. 

16mo, mor., 1 50 

MOLITOR and BEARD Manual for Resident Engineers 16mo, 1 00 

NAGLE Field Manual for Railroad Engineers ; . . . 16mo, mor., 3 00 

ORROCK Railroad Structures and Estimates Svo, *3 00 

PHILBRICK Field Manual for Engineers 16mo, mor., 3 00 

RAYMOND Elements of Railroad Engineering Svo, 3 50 

Railroad Engineer's Field Book (In Preparation.) 

Railroad Field Geometry lomo, mor., 2 00 

ROBERTS Track Formulae and Tables 16mo, mor., 3 00 , 

SEARLES Field Engineering 16mo, mor., 3 00 

Railroad Spiral 16mo, mor., 1 50 

TAYLOR Prismoidal Formulae and Earthwork Svo, 1 50 

WEBB Economics of Railroad Construction Small Svo, 2 50 

Railroad Construction 16mo, mor., 5 00 

WELLINGTON Economic Theory of the Location of Railways. . . .Small Svo, 5 00 

WILSON Elements of Railroad-Track and Construction 12mo, 2 00 


BARR and WOOD Kinematics of Machinery 8vo, $2 50 

BARTLETT Mechanical Drawing. Third Edition 8vo, *3 00 

" t " Abridgment of the Second Edition. .. .8vo, *1 50 

and JOHNSON Engineering Descriptive Geometry 8vo, *1 50 

BISHOP Structural Details of Hip and Valley Rafters. . . .Oblong large 8vo, *1 75 

BLESSING and DARLING Descriptive Geometry 8vo, *1 50 

Elements of Drawing 8vo, *1 50 

COOLIDGE Manual of Drawing 8vo, paper, 1 00 

and FREEMAN Elements of General Drafting for Mechanical Engineers 

Oblong 4to, 2 50 

DURLEY Kinematics of Machines 8vo, 4 00 

EMCH Introduction to Protective Geometry and its Application 8vo, 2 50 

FRENCH and IVES Stereotomy 8vo, 2 50 

HILL Text-book on Shades and Shadows, and Perspective 8vo, 2 00 

JAMISON Advanced Mechanical Drawing 8vo, 2 00 

Elements of Mechanical Drawing Svo, 2 50 

JONES Machine Design: 

Part I. Kinematics of Machinery Svo, 1 50 

Part II. Form, Strength, and Proportions of Parts Svo, 3 00 

KIMBALL and BARR Machine Design Svo, *3 00 

MAcCoRD Elements of Descriptive Geometry Svo, 3 00 

Kinematics; or, Practical Mechanism Svo, 5 00 

Mechanical Drawing 4to, 4 CO 

Velocity Diagrams Svo, 1 50 

McLEOD Descriptive Geometry Small Svo, 1 50 

MAHAN Descriptive Geometry and Stone-cutting Svo, *1 50 

Industrial Drawing. (THOMPSON.) Svo, 3 50 

MOVER Descriptive Geometry Svo, 2 00 

REED Topographical Drawing and Sketching 4to, 5 00 

REID Mechanical Drawing. (Elementary and Advanced.) Svo, *2 00 

Text-book of Mechanical Drawing and Elementary Machine Design. Svo, 3 00 

ROBINSON Principles of Mechanism Svo, 3 00 

SCHWAMB and MERRILL Elements of Mechanism Svo, 3 00 

SMITH (A. W.) and MARX Machine Design Svo, 3 00 

(R. S.) Manual of Topographical Drawing. (McMiLLAX.) Svo, 2 50 

TITSWORTH Elements of Mechanical Drawing Oblong large Svo, *1 25 

WARREN Elements of Descriptive Geometry, Shadows, and Perspective. Svo, 3 50 

Elements of Machine Construction and Drawing Svo, 7 50 

Elements of Plane and Solid Free-hand Geometrical Drawing. . . . 12mo, 1 00 

General Problems of Shades and Shadows Svo, 3 00 

Manual of Elementary Problems in the Linear Perspective of Forms and 

Shadows 12mo, 1 00 

Manual of Elementary Projection Drawing 12mo, 1 50 

Plane Problems in Elementary Geometry 12mo, 1 25 

WEISBACH Kinematics and Power of Transmission. (HERRMANN and 

KLEIN.) Svo, 5 00 

WiLSON*(H. M.) Topographic, Trigonometric and Geodetic Surveying. Svo, 3 50 

(V. T.) Descriptive Geometry Svo, *1 50 

Free-hand Lettering Svo, 1 00 

Free-hand Perspective Svo, 2 50 

WOOLF Elementary Course in Descriptive Geometry Large Svo, 3 00 


ABEGG Theory of Electrolytic Dissociation. (VON ENDE.) 12mo, *1 25 

ANDREWS Hand-book for Street Railway Engineers 3 X5 inches, mor., 1 25 

ANTHONY and BALL Lecture-notes on the Theory of Electrical Measure- 
ments 12mo, 1 00 

and BRACKETT Text-book of Physics. (MAGIE.) Small Svo, 3 00 

BENJAMIN History of Electricity Svo, 3 00 

BETTS Lead Refining and Electrolysis Svo, 4 00 

BURGESS and LE CHATELIER Measurement of High Temperatures. Third 

Edition Svo, *4 00 

CLASSEN Quantitative Analysis by Electrolvsis. (HALL.) (In Press.) 


COLLINS Manual of W4reless Telegraphy and Telephony 12mo,*$l 50 

CREHORE and SQUIER Polarizing Photo-chronograph 8vo, 3 00 

DANNEEL Electrochemistry. (MERRIAM.) 12mo, *1 25 

DAWSON " Engineering " and Electric Traction Pocket-book. . 16mo, mor., 5 00 
DOLEZALEK Theory of the Lead Accumulator (Storage Battery). (VON 

ENDE.) 12mo, 2 50 

DUHEM Thermodynamics and Chemistry. (BURGESS.) 8vo, 4 00 

FLATHER Dynamometers, and the Measurement of Power 12mo, 3 00 

GETMAN Introduction to Physical Science 12mo, *1 50 

GILBERT De Magnete. (MOTTELAY.) 8vo, 2 50 

HANCHETT Alternating Currents 12mo, *1 00 

HERING Ready Reference Tables (Conversion Factors) IGmo, mor., 2 50 

HOBART and ELLIS High-speed Dynamo Electric Machinery 8vo, *6 00 

HOLMAN Precision of Measurements ' 8vo, 2 00 

Telescope- Mirror-scale Method, Adjustments, and Tests. . . .Large 8vo, 75 
HUTCHINSON High-Efficiency Electrical Illuminants and Illumination. 

Small 8vo, *2 50 

JONES Electric Ignition for Combustion Motors 8vo, *4 00 

KARAPETOFF Experimental Electrical Engineering: 

Vol. I Svo, *3 50 

Vol. II 8vo, *2 50 

KINZBRUNNER Testing of Continuous-current Machines Svo, 2 00 

KOCH Mathematics of Applied Electricity Small Svo, *3 00 

LANDAUER Spectrum Analysis. (TINGLE.) 8vo, 3 00 

LAUFFER Electrical Injuries Ifimo, *0 50 

LOB Electrochemistry of Organic Compounds. (LORENZ.) Svo, 3 00 

LYNDON Development and Electrical Distribution of Water Power.. . .8vo, *3 00 
LYONS Treatise on Electromagnetic Phenomena. Vols. I and II, Svo, each, *6 00 

MARTIN Measurement of Induction Shocks 12mo, *1 25 

MICHIE Elements of Wave Motion Relating to Sound and Light Svo, *4 00 

MORGAN Physical Chemistry for Electrical Engineers 12mo, *1 50 

NORRIS Introduction to the Study of Electrical Engineering Svo, *2 50 

PARSHALL and HOBART Electric Machine Design 4to, half mor., *12 50 

REAGAN Locomotives: Simple, Compound, and Electric Small Svo, 3 50 

RODENHAUSER and SCHOENAWA Electric Furnaces in the Iron and Steel 

Industry (VoM BAUR.) (In Press.) 

ROSENBERG Electrical Engineering. (HALDANE GEE KINZBRUNNER.) .Svo, *2 00 
RYAN Design of Electrical Machinery: 

Vol. I. Direct Current Dynamos Svo, *1 50 

Vol. II. Alternating Current Transformers Svo, *1 50 

Vol. III. Alternators, Synchronous Motors, and Rotary Converters. 

Svo, *1 50 
SCHAPPER Laboratory Guide for Students in Physical Chemistry. . . . 12mo, 1 00 

TILLMAN Elementary Lessons in Heat Svo, *1 50 

TIMBIE Answers to Problems in Elements of Electricity 12mo, Paper, *0 25 

Elements of Electricity Small Svo, *2 00 

Essentials of Electricity 12mo, *1 25 

TORY and PITCHER Manual of Laboratory Physics Small Svo, 2 00 

ULKE Modern Electrolytic Copper Refining Svo, 3 00 

WATERS Commercial Dynamo Design Svo, *2 00 


BRENNAN Hand-book of Useful Legal Information for Business Men. 

16mo, mor., *5 00 

DAVIS Elements of Law Svo, *2 50 

Treatise on the Military Law of United States Svo, *7 00 

DUDLEY Military Law and the Procedure of Courts-martial. . .Small Svo, *2 50 


WAIT Engineering and Architectural Jurisprudence Svo, 6 00 

Sheep, 6 50 

Law of Contracts Svo, 3 00 

Law of Operations Preliminary to Construction in Engineering and 

Architecture Svo, 5 00 

Sheep, 5 50 


BAKER Elliptic Functions Svo, $1 50 

BRIGGS Elements of Plane Analytic Geometry. (BociiER.) 12mo, 1 00 

BUCHANAN Plane and Spherical Trigonometry 8vo, *1 00 

BYERLY Harmonic Functions 8vo, 1 00 

CHANDLER Elements of the Infinitesimal Calculus 12mo, 2 00 

COFFIN Vector Analysis 12mo, *2 50 

COMPTON Manual of Logarithmic Computations 12mo, 1 50 

DICKSON College Algebra Small 8vo, *1 50 

Introduction to the Theory of Algebraic Equations Small 8vo, *1 25 

EMCH Introduction to Projective Geometry and its Application 8vo, 2 50 

FISKE Functions of a Complex Variable 8vo, 1 00 

HALSTED Elementary Synthetic Geometry 8vo, 1 50 

Elements of Geometry 8vo, 1 75 

Rational Geometry 12mo, *1 50 

Synthetic Projective Geometry 8vo, 1 00 

HANCOCK Lectures on the Theory of Elliptic Functions 8vo, *5 00 

HYDE Grassmann's Space Analysis 8vo, 1 00 

JOHNSON (J. B.) Three-place Logarithmic Tables: Vest-pocket size, paper, *0 15 

100 copies, *5 00 

Mounted on heavy cardboard, 8 X10 inches, *0 25 

10 copies, *2 00 
(W. W.) Abridged Editions of Differential and Integral Calculus. 

Small 8vo, 1 vol.. 2 50 

Curve Tracing in Cartesian Co-ordinates . 12mo, 1 00 

Differential Equations 8vo, 1 00 

Elementary Treatise on Differential Calculus Small 8vo, 1 50 

Elementary Treatise on the Integral Calculus Small 8vo, 1 50 

Theoretical Mechanics 12mo, *3 00 

Theory of Errors and the Method of Least Squares 12mo, 1 50 

Treatise on Differential Calculus Small 8vo, 3 00 

Treatise on the Integral Oalculus Small 8vo, 3 00 

Treatise on Ordinary and Partial Differential Equations .... Small 8vo, 3 50 
KARAPETOFF Engineering Applications of Higher Mathematics: 

Part I. Problems on Machine Design '. .Small 8vo, *0 75 

KOCH Mathematics of Applied Electricity Small 8vo, *3 00 

LAPLACE Philosophical Essay on Probabilities. (TRUScoxr and EMORY.) 

12mo. 2 00 
LE MESSURIER Key to Professor W. W. Johnson's Differential Equations. 

Small Svo, *1 75 

LUDLOW Logarithmic and Trigonometric Tables Svo, *1 00 

and BASS Elements of Trigonometry and Logarithmic and Other 

Tables Svo, *3 00 

Trigonometry and Tables published separately Each, *2 00 

MACFARLANE Vector Analysis and Quaternions Svo, 1 00 

McMAHON Hyperbolic Functions Svo, 1 00 

MANNING Irrational Numbers and their Representation by Sequences and 

Series 12mo, 1 25 

MARSH Industrial Mathematics Small Svo, *2 00 


ROBERT S. WOODWARD Octavo, each, 1 00 

No. 1. History of Modern Mathematics, by DAVID EUGENE SMITH. 

No. 2. Synthetic Projective Geometry, by GEORGE BRUCE HALSTED. 

No. 3. Determinants, by LAENAS GIFFORD WELD. 

No. 4. Hyperbolic Functions, by JAMES McMAHON. 

No. 5, Harmonic Functions, by WILLIAM E. BYERLY. 

No. 6. Grassmann's Space Analysis, by EDWARD W. HYDE. 

No. 7. Probability and Theory of Errors, by ROBERT S. WOODWARD. 

No. 8. Vector Analysis and Quaternions, by ALEXANDER MACFARLANE. 

No. 9. Differential Equations, by WILLIAM WOOLSEY JOHNSON. 

No. 10. The Solution of Equations, by MANSFIELD MERRIMAN. 

No. 11. Functions of a Complex Variable, by THOMAS S. FISKE. 

MAURER Technical Mechanics Svo, 4 00 

MERRIMAN Method of Least Squares Svo, 2 00 

Solution of Equations Svo, 1 00 

MORITZ Elements of Plane Trigonometry Svo, *2 00 

High School Edition Small Svo, *1 00 


RICE and JOHNSON Differential and Integral Calculus. 2 vols. in one. 

Small 8vo, $1*50 

Elementary Treatise on the Differential Calculus Small 8vo, 3 00 

SMITH History of Modern Mathematics 8vo, 1 00 

VEBLEN and LENNES Introduction to the Real Infinestimal Analysis of One 

Variable 8vo, *2 00 

WATERBURY Vest Pocket Hand-book of Mathematics for Engineers. 

2| X5| inches, mor., *1 00 

Enlarged Edition, Including Tables mor., *1 50 

WELD Determinants 8vo, 1 00 

WOOD Elements of Co-ordinate Geometry . .8vo, 2 00 

WOODWARD Probability and Theory of Errors 8vo, 1 00 



BACON Forge Practice 12mo, 1 50 

BALDWIN Steam Heating for Buildings 12mo, 2 50 

BARR and WOOD Kinematics of Machinery 8vo, 2 50 

BARTLETT Mechanical Drawing. Third Edition 8vo, 3 00 

Abridgment of the Second Edition.. . .8vo, *1 50 

BURR Ancient and Modern Engineering and the Isthmian Canal 8vo, *3 50 

CARPENTER Heating and Ventilating Buildings 8vo, 4 00 

and DIEDERICHS Experimental Engineering 8vo, *6 00 

CLERK The Gas, Petrol and Oil Engine 8vo, *4 00 

COMPTON First Lessons in Metal Working 12mo, 1 50 

and DE GROODT Speed Lathe 12mo, 1 50 

COOLIDGE Manual of Drawing 8vo, paper, 1 00 

and FREEMAN Elements of General Drafting for Mechanical Engi- 
neers Oblong 4to, 2 50 

CROMWELL Treatise on Belts and Pulleys 12mo, 1 50 

Treatise on Toothed Gearing 12mo, 1 50 

DINGEY Machinery Pattern Making 12mo, 2 00 

DURLEY Kinematics of Machines 8vo, 4 00 

FLANDERS Gear-cutting Machinery Small 8vo, 3 00 

FLATHER Dynamometers and the Measurement of Power 12mo, 3 00 

Rope Driving 12mo, 2 00 

FULLER and JOHNSTON Applied Mechanics: 

Vol. I. Theory of Statics and Kinetics (In Press.) 

Vol. II. Strength of Materials (In Preparation.) 

GILL Gas and Fuel Analysis for Engineers 12mo, 1 25 

Goss Locomotive Sparks 1 8vo, 2 00 

GREENE Elements of Heating and Ventilation Svo, *2 50 

Pumping Machinery Svo, *4 00 

HERING Ready Reference Tables (Conversion Factors) IGmo, mor., 2 50 

HOBART and ELLIS High Speed Dynamo Electric Machinery Svo, *6 00 

HUTTON Gas Engine Svo, 5 00 

JAMISON Advanced Mechanical Drawing Svo, 2 00 

Elements of Mechanical Drawing. Svo, 2 50 

JONES Gas Engine Svo, 4 00 

Machine Design: 

Part I. Kinematics of Machinery Svo, 1 50 

Part II. Form, Strength, and Proportions of Parts. .' Svo, 3 00 

KAUP Machine Shop Practice Small Svo, *1 25 

KENT Mechanical Engineers' Pocket-Book IGmo, mor., *5 00 

KERR Power and Power Transmission Svo, 2 00 

KIMBALL and BARR Machine Design Svo, *3 00 

KING Elements of the Mechanics of Materials and of Power of Trans- 
mission Svo, *2 50 

LANZA Dynamics of Machinery Svo, *2 50 

LEONARD Machine Shop Tools and Methods Svo, 4 00 

LEVIN Modern Gas Bngine and the Gas Producer Svo, *4 00 

LORENZ Modern Refrigerating Machinery. (POPE, HAVEN, and DEAN.) 

Svo, *4 00 


MACCORD Kinematics; or, Practical Mechanism 8vo, $5 00 

Mechanical Drawing 4to, 4 00 

Velocity Diagrams 8vo, 1 50 

MACFARLAND Standard Reduction Factors for Gases 8vo, 1 50 

MAHAN Industrial Drawing. (THOMPSON.) 8vo, 3 50 

MEHRTENS Gas Engine Theory and Design Small 8vo, 2 50 

OBERG Handbook of Small Tools Small 8vo, 2 50 

PARSHALL and HOBART Electric Machine Design. . .Small 4to, half leather, *12 50 
PEELE Compressed Air Plant. Second Edition, Revised and Enlarged. 8vo, *3 50 

POOLE Calorific Power of Fuels 8vo, 3 00 

PORTER Engineering Reminiscences, 1855 to 1882 8vo, *3 00 

REID Mechanical Drawing. (Elementary and Advanced.) 8vo, *2 00 

Text-book of Mechanical Drawing and Elementary Machine Design. Svo, 3 00 

RICHARDS Compressed Air 12mo, 1 50 

ROBINSON Principles of Mechanism Svo, 3 00 

SCHWAMB and MERRILL Elements of Mechanism Svo, 3 00 

SMITH (O.) Press-working of Metals Svo, 3 00 

(A. W.) and MARX Machine Design Svo, 3 00 

SOREL Carbureting and Combustion in Alcohol Engines. (WOODWARD and 

PRESTON.) Small Svo, 3 00 

STONE Practical Testing of Gas and Gas Meters Svo, 3 50 

THURSTON Animal as a Machine and Prime Motor, and the Laws of 

Energetics 12mo, 1 00 

Treatise on Friction and Lost Work in Machinery and Mill Work. .Svo, 3 00 

TILLSON Complete Automobile Instructor 16mo, *1 50 

TITSWORTH Elements of Mechanical Drawing Oblong Svo, *1 25 

WARREN Elements of Machine Construction and Drawing Svo, 7 50 

WATERBURY Vest Pocket Hand-book of Mathematics for Engineers. 

2 1 X5 inches, mor., *1 00 

Enlarged Edition, Including Tables mor., *1 50 

WEISBACH Kinematics and the Power of Transmission. (HERRMANN 

KLEIN.) Svo, 5 00 

Machinery of Transmission and Governors. (HERRMANN KLEIN.) .Svo, 5 00 
WOOD Turbines Svo, 2 50 


BOTTLER German and American Varnish Making. (SABIN.) . . .Small Svo, *3 50 

BURR Elasticity and Resistance of the Materials of Engineering Svo, 7 50 

CHURCH Mechanics of Engineering Svo, 6 00 

Mechanics of Solids (Being Parts I, II, III of Mechanics of Engineering). 

Svo, 4 50 
FULLER and JOHNSTON Applied Mechanics: 

Vol. I. Theory of Statics and Kinetics (In Press.) 

Vol. II. Strength of Materials (In Preparation.) 

GREENE Structural Mechanics Svo, *2 50 

HOLLEY Analysis of Paint and Varnish Products Small Svo, *2 50 

Lead and Zinc Pigments Small Svo, *3 00 

JOHNSON (C. M.) Rapid Methods for the Chemical Analysis of Special 

Steels, Steel-making Alloys and Graphite Small Svo, 3 00 

(J. B.) Materials of Construction Svo, 6 00 

KEEP Cast Iron Svo, 2 50 

KING Elements of the Mechanics of Materials and of Power of Trans- 
mission Svo, *2 50 

LANZA Applied Mechanics Svo, 7 50 

LOWE Paints for Steel Structures 12mo, 1 00 

MAIRE Modern Pigments and their Vehicles 12mo, 2 00 

MARTIN Text-Book of Mechanics: 

Vol. I. Statics 12mo, *1 25 

Vol. II. Kinematics and Kinetics 12mo, *1 50 

Vol. III. Mechanics of Materials 12mo, *1 50 

Vol. IV. Applied Statics 12mo, *1 50 

MAURER Technical Mechanics Svo, 4 00 

MERRIMAN Mechanics of Materials Svo, 5 00 

Strength of Materials 12mo, *1 00 

METCALF Steel. A Manual for Steel-users 12mo, 2 00 

MURDOCK Strength of Materials 12mo, *2 00 


SARIN Industrial and Artistic Technology of Paint and Varnish 8vo, $3 00 

SMITH (A. W.) Materials of Machines 12mo, 1 00 

(H. E.) Strength of Material 12mo, *1 25 

THURSTON Materials of Engineering 3 vols., 8vo, 8 00 

Part I. Non-metallic Materials of Engineering 8vo, 2 00 

Part II. Iron and Steel 8vo, 3 50 

Part III. A Treatise on Brasses, Bronzes, and Other Alloys and their 

Constituents 8vo, 2 50 

WATERBURY Laboratory Manual for Testing Materials of Construction. 

12mo, *1 50 

WOOD (DE V.) Elements of Analytical Mechanics 8vo, 3 00 

Treatise on the Resistance of Materials and an Appendix on the Preser- 
vation of Timber 8vo, 2 00 

(M. P.) Rustless Coatings. Corrosion and Electrolysis of Iron and 

Steel 8vo, 4 00 


ABRAHAM Steam Economy in the Sugar Factory. (BAYLE.) . . . (In Press.) 
BERRY Temperature-entropy Diagram. Third Edition Revised and En- 
larged ; 12mo, , 2 50 

CARNOT Reflections on the Motive Power of Heat. (THURSTON.).. .12mb, 1 50 

CHASE Art of Pattern Making 12mo, 2 50 

CREIOHTON Steam-engine and other Heat Motors 8vo, 5 00 

DAWSON ""Engineering " and Electric Traction Pocket-book. . .IGrno, mor., 5 00 

GEBHARDT Steam Power Plant Engineering 8vo, *6 00 

Goss Locomotive Performance 8vo, 5 00 

HEMENWAY Indicator Practice and Steam-engine Economy 12mo, 2 00 

HIRSHKELD and BARNARD Heat Power Engineering 8vo, *5 00 

HUTTON Heat and Heat-engines 8vo, 5 00 

Mechanical Engineering of Power Plants 8vo, 5 00 

KENT Steam Boiler Economy 8vo, 4 00 

KING Steam Engineering (In Press.) 

KNEASS Practice and Theory of the Injector 8vo, 1 50 

MACCORD Slide-valves 8vo, 2 00 

MEYER Modern Locomotive Construction 4to, 10 00 

MILLER, BERRY, and RILEY Problems in Thermodynamics.. . .8vo, paper, 75 

MOVER Steam Turbines 8vo, 4 00 

PEABODY Manual of the Steam-engine Indicator 12mo, 1 50 

Tables of the Properties of Steam and Other Vapors and Temperature- 
Entropy Table 8vo, 1 00 

Thermodynamics of the Steam-engine and Other Heat-engines 8vo, 5 00 

Thermodynamics of the Steam Turbine 8vo, *3 00 

Valve-gears for Steam-engines 8vo, 2 50 

and MILLER Steam-boilers 8vo, 4 00 

PERKINS Introduction to General Thermodynamics 12mo, *1 50 

PUPIN Thermodynamics of Reversible Cycles in Gases and Saturated 

Vapors. (OSTERBERG.) 12mo, 1 25 

REAGAN Locomotives: Simple, Compound, and Electric. New Edition. 

Small 8vo, 3 50 

SINCLAIR Locomotive Engine Running and Management 12mo, 2 00 

SMART Handbook of Engineering Laboratory Practice 12mo, 2 50 

SNOW Steam-boiler Practice 8vo, 3 00 

SPANGLER Notes on Thermodynamics 12mo, 1 00 

Valve-gears 8vo, 2 50 

GREENE, and MARSHALL Elements of Steam-engineering 8vo, 3 00 

THOMAS Steam-turbines 8vo, 4 00 

THURSTON Manual of Steam-boilers, their Designs, Construction, and 

Operation 8vo, 5 00 

Manual of the Steam-engine 2 vols., 8vo, 10 00 

Part I. History, Structure, and Theory 8vo, 6 00 

Part II. Design, Construction, and Operation 8vo, 6 00 

WEHRENFENNIG Analysis and Softening of Boiler Feed-water. (PATTER- 
SON.) 8vo, 4 00 

WEISBACH Heat, Steam, and Steam-engines. (Du Bois.) 8vo, 5 00 

WHITHAM Steam-engine Design 8vo, 5 00 

WOOD Thermodynamics, Heat Motors, and Refrigerating Machines. . .8vo, 4 00 



CHURCH Mechanics of Engineering .... 8vo, $6 00 

Mechanics of Solids (Being Parts I, 11, III of Mechanics of Engineering). 

8vo, 4 50 

Mechanics of Fluids (Being Part IV of Mechanics of Engineering) .8vo, 3 00 

Mechanics of Internal Work 8vo, *1 50 

Notes and Examples in Mechanics 8vo, 2 00 

DANA Text-book of Elementary Mechanics for Colleges and Schools . .12mo, 1 50 
Du Bois Elementary Principles of Mechanics: 

Vol. I. Kinematics 8vo, 3 50 

Vol. II. Statics 8vo, 4 00 

Mechanics of Engineering. Vol. I Small 4to, 7 50 

Vol. II Small 4to, 10 00 

FULLER and JOHNSTON Applied Mechanics: 

Vol. I. Theory of Statics and Kinetics (In Press.} 

Vol. II. Strength of Materials (In Preparation.) 

GREENE Structural Mechanics 8vo, *2 50 

HARTMANN Elementary Mechanics for Engineering Students 12mo, *1 25 

JAMES Kinematics of a Point and the Rational Mechanics of a Particle. 

Small 8vo, 2 00 

JOHNSON' (W. W.) Theoretical Mechanics 12mo, *3 00 

KING Elements of the Mechanics of Materials and of Power of Trans- 
mission 8vo, *2 50 

KOTTCAMP Exercises for the Applied Mechanics Laboratory, Loose Leaf 

Laboratory Manual. Oblong 4to, paper, *1 00 

LANZA Applied Mechanics 8vo, 7 50 

MARTIN Text Book of Mechanics: 

Vol. I. Statics 12mo, *1 25 

Vol. II. Kinematics and Kinetics 12mo, *1 50 

Vol. III. Mechanics of Materials 12mo, *1 50 

Vol. IV. Applied Statics 12mo, *1 50 

MAURER Technical Mechanics 8vo, 4 00 

MERRIMAN Elements of Mechanics 12mo, *1 00 

Mechanics of Materials 8vo, 5 00 

MICHIE Elements of Analytical Mechanics Svo, *4 00 

ROBINSON Principles of Mechanism Svo, 3 00 

SANBORN Mechanics Problems Small Svo, *1 50 

SCHWAMB and MERRILL Elements of Mechanism Svo, 3 00 

WOOD Elements of Analytical Mechanics Svo, 3 00 

Principles of Elementary Mechanics 12mo, 1 25 


ABDERHALDEN Physiological Chemistry in Thirty Lectures. (HALL and 

DEFREN.) Svo, *5 00 

VON BEHRING Suppression of Tuberculosis. (Bolduan.) 12mo, 1 00 

BOLDUAN Immune Sera 12mo, *1 50 

BORDET Studies in Immunity. (GAY.) Svo, 6 00 

CHAPIN The Sources and Modes of Infection Small Svo, *3 00 

COHNHEIM Enzymes 12mo, * 1 50 

DAVENPORT Statistical Methods with Special Reference to Biological Varia- 
tions 16mo, mor., 1 50 

EFFRONT Enzymes and Their Applications. (PRESCOTT.) Svo, 3 00 

EHRLICH Studies on Immunity. (BOLDUAN.) Svo, G 00 

EULER General Chemistry of the Enzymes. (PoPE.) Svo, *3 00 

FISCHER Nephritis Small Svo, *2 50 

Oedema Svo, *2 00 

Physiology of Alimentation .' Small Svo, *2 00 

DE FURSAC Manual of Psychiatry. (ROSANOFF and COLLINS.) .Small Svo, *2 50 

FULLER Qualitative Analysis of Medicinal Preparations 12mo, *1 50 

HAMMARSTEN Text-book on Physiological Chemistry. (MANDEL.).. . .Svo, *4 00 
JACKSON Directions for Laboratory Work in Physiological Chemistry . Svo, 1 25 

LASSAR-COHN Praxis of Urinary Analysis. (LORENZ.) 12mo, 1 00 

LAUFFER Electrical Injuries 16mo, *0 50 

MANDEL Hand-book for the Bio-Chemical Laboratory 12mo, 1 50 


MARTIN Measurement of Induction Shocks 12mo, *1 25 

NELSON Analysis of Drugs and Medicines 12mo, *3 00 

PAULI Physical Chemistry in the Service of Medicine. (FISCHER.) . . . 12mo, *1 25 
Pozzi-EscoT Toxins and Venoms and their Antibodies. (CoHN.) . . . 12mo, *1 00 

ROSTOSKI Serum Diagnosis. (BOLDUAN.) 12mo, 1 00 

RUDDIMAN Incompatibilities in Prescriptions 8vo, 2 00 

Whys in Pharmacy 12mo, 1 00 

SALKOWSKI Physiological and Pathological Chemistry. (ORNDORFF.) . 8vo, 2 50 

SATTERLEE Outlines of Human Embryology 12mo, *1 25 

SMITH Lecture Notes on Chemistry for Dental Students 8vo, *2 50 

WHIFFLE Typhoid Fever Small 8vo, *3 00 

Woor>HULL Military Hygiene for Officers of the Line Small 8vo, *1 50 

Personal Hygiene 12mo, *1 00 

WORCESTER and ATKINSON Small Hospitals Establishment and Mainte- 
nance, and Suggestions for Hospital Architecture, with Plans for a 
Smalt Hospital 12mo, 1 25 


BETTS Lead Refining by Electrolysis 8vo, 4 00 

BOLLAND Encyclopedia of Founding and Dictionary of Foundry Terms 

used in the Practice of Moulding 12mo, 3 00 

Iron Founder 12mo, 2 50 

BORCHERS Metallurgy. (HALL and HAYWARD.) 8vo, *3 00 

BURGESS and LE CHATELIEK Measurement of High Temperatures. Third 

Edition 8vo, *4~00 

DOUGLAS Untechnical Addresses on Technical Subjects 12mo, 1 00 

GOESEL Minerals andjMetals: A Reference Book 16mo, mor., 3 00 

ILES Lead-smelting 12mo, *2 50 

JOHNSON Rapid Methods for the Chemical Analysis of Special Steels, 

Steel-making Alloys and Graphite Large, 12mo, 3 00 

KEEP Cast Iron 8vo, 2 50 

METCALF Steel. A Manual for Steel-users 12mo, 2 00 

MINET Production of Aluminum and its Industrial Use. (WALDO.). . 12mo, 2 50 

PALMER Foundry Practice Small Svo, *2 00 

PRICE and MEADE Technical Analysis of Brass 12mo, *2 00 

RODENHAUSER and SCHOENAWA Electric Furnaces in the Iron and Steel 

Industry. (VoM BAUR.) (In Press) 

R UER Elements of Metallography. (MATHEWSON.) Svo, *3 00 

SMITH Materials of Machines 12mo, 1 00 

TATE and STONE Foundry Practice 12mo, 2 00 

THURSTON Materials of Engineering. In Three Parts Svo, 8 00 

Part I. Non-metallic Materials of Engineering, see Civil Engineering, 

page 9. 

Part II. Iron and Steel Svo, 3 50 

Part III. A Treatise on Brasses, Bronzes, and Other Alloys"and Their 

Constituents Svo, 2 50 

UlKE Modern Electrolytic Copper Refining Svo, 3 00 

WEST American Foundry Practice 12mo, 2 50 

Moulders' Text Book 12mo, 2 50 



BERNADOU Smokeless Powder, Nitro-cellulose, and the Theory of the Cellu- 
lose Molecule 12mo, 2 50 

CHASE Art of Pattern Making 12mo, 2 50 

Screw Propellers and Marine Propulsion Svo, 3 00 

CLOKE Enlisted Specialists' Examiner Svo, *2 00 

Gunner's Examiner Svo, *1 50 

CRAIG Azimuth 4to, 3 50 

CREHORE and SQUIER Polarizing Photo-chronograph Svo, 3 00 

DAVIS Elements of Law Svo, *2 50 

Treatise on the Military Law of United States Svo, *7 00 


DUDLEY Military Law and the Procedure of Courts-martial. . ..Small 8vo, *2 50 

DURAND Resistance and Propulsion of Ships 8vo, 5 00 

DYER Handbook of Light Artillery 12mo, *3 00 

EISSLER Modern High Explosives 8vo, 4 00 

FIEBEGER Text-book on Field Fortification Small 8vo, *2 00 

HAMILTON and BOND The Gunner's Catechism. . . 18mo, 1 00 

HOFF Elementary Naval Tactics 8vo, *1 50 

INGALLS Handbook of Problems in Direct Fire 8vo, 4 00 

Interior Ballistics 8vo, *3 00 

LISSAK Ordnance and Gunnery 8vo, *6 00 

LUDLOW Logarithmic and Trigonometric Tables 8vo, *1 00 

LYONS Treatise on Electromagnetic Phenomena. Vols. I. and II., 8vo, each, *6 00 

MAHAN Permanent Fortifications. (MERCUR) 8vo, half mor., *7 50 


MERCUR Attack of Fortified Places 12mo,$*2 00 

Elements of the Art of War 8vo, *4 00 

NIXON Adjutants' Manual 24mo, 1 00 

PEABODY Naval Architecture 8vo, 7 50 

Propellers 8vo, 1 25 

PHELPS Practical Marine Surveying 8vo, *2 50 

PUTNAM Nautical Charts 8vo, 2 00 

RUST Ex-meridian Altitude, Azimuth and Star-Finding Tables 8vo, 5 00 

SELKIRK Catechism of Manual of Guard Duty 24mo, *0 50 

SHARPE Art of Subsisting Armies in War 18mo, mor., 1 50 

TAYLOR Speed and Power of Ships. 2 vols. Text 8vo, plates oblong 4to, *7 50 
TUPES and POOLE Manual of Bayonet Exercise and Musketry Fencing. 

24mo, leather, *0 50 

WEAVER Military Explosives 8vo, *3 00 

WOODHULL Military Hygiene for Officers of the Line Small 8vo, *1 50 


BROWNING Introduction to Rarer Elements 8vo, *1 50 

BRUSH Manual of Determinative Mineralogy. (PHNFIELD.) 8vo, 4 00 

BUTLER Pocket Hand-book of Blowpipe Analysis 16mo, *0 75 

Pocket Hand-book of Minerals 16mo, mor., 3 00 

CHESTER Catalogue of Minerals 8vo, paper, 1 00 

Cloth, 1 25 

CRANE Gold and Silver 8vo, *5 00 

DANA First Appendix to Dana's New " System of Mineralogy." . Large 8vo, 1 00 
Second Appendix to Dana's New " System of Mineralogy." Large 8vo, 1 50 

Manual of Mineralogy. (FoRD.) 12mo, *2 00 

Minerals, and How to Study Them 12mo, 1 50 

System of Mineralogy Large 8vo, half leather, 12 50 

Text-book of Mineralogy 8vo, 4 00 

DOUGLAS Untechnical Addresses on Technical Subjects 12mo, 1 00 

EAKLE Mineral Tables 8vo, 1 25 

ECKEL Building Stones and Clays 8vo, *3 00 

GOESEL Minerals and Metals: A Reference Book 16mo, mor., 3 00 

GROTH The Optical Properties of Crystals. (JACKSON.) 8vo, *3 50 

Introduction to Chemical Crystallography. (MARSHALL.) 12mo, 1 25 

HAYES Handbook for Field Geologists IGmo, mor., *1 50 

IDDINGS Igneous Rocks 8vo, 5 00 

Rock Minerals 8vo, 5 00 

JOHANNSEN Determination of Rock-forming Minerals in Thin Sections, 8vo, 

With Thumb Index, 5 00 

LEWIS Determinative Mineralogy Small 8vo, *1 50 

MARTIN Laboratory Guide to Qualitative Analysis with the Blowpipe. 12mo, *0 60 

MERRILL Non-metallic Minerals: Their Occurrence and Uses 8vo, 4 00 

, Stones for Building and Decoration 8vo, 5 00 

PENFIELD Notes on Determinative Mineralogy and Record of Mineral 

Tests 8vo, paper, *0 50 

Tables of Minerals, Including the Use of Minerals and Statistics of 

Domestic Production 8vo, 1 00 

PIRSSON Rocks and Rock Minerals 12mo, *2 50 

RICHARDS Synopsis of Mineral Character 12mo, mor., *1 25 


RIES Building Stones and Clay Products 8vo,*$3 00 

Clays: Their Occurrendfe, Properties and Uses 8vo, *5 00 

and LEIGHTON History of the Clay-working Industry of the United 

States 8vo, *2 50 

ROWE Practical Mineralogy Simplified 12mo, *1 25 

TILLMAN Text-book of Important Minerals and Rocks 8vo, *2 00 

WASHINGTON Manual of the Chemical Analysis of Rocks 8vo, 2 00 


BEARD Mine Gases and Explosions Small Svo, *3 00 

BRUNSWIG Explosives. (MUNROE and KIBLER.) Ready Fall, 1912 

CRANE Gold and Silver 8vo, *5 00 

Index of Mining Engineering Literature, Vol. I Svo, *4 00 

Svo, mor., *5 QO 

Vol. II Svo, *3 00 

Svo, mor., *4 00 

Ore Mining Methods Svo, *3 00 

DANA and SAUNDERS Rock Drilling Svo, *4 00 

DOUGLAS Untechnical Addresses on Technical Subjects 12mo, 1 00 

EISSLER Modern High Explosives Svo, 4 00 

GILBERT, WIGHTMAN and SAUNDERS Subways and Tunnels of New York. 

Svo, *4 00 

GOESEL Minerals and Metals: A Reference Book IGmo, mor., 3 00 

IHLSENG Manual of Mining Svo, 5 00 

ILES Lead Smelting 12mo, *2 50 

PEELE Compressed Air Plant Svo, *3 50 

RIEMER Shaft Sinking under Difficult Conditions. (CORDING and PEELE.) 

Svo, 3 00 

WEAVER Military Explosives Svo, *3 00 

WILSON Hydraulic and Placer Mining 12mo, 2 50 

Treatise on Practical and Theoretical Mine Ventilation 12mo, 1 25 



Hartford Meeting, 1900 Svo, 3 00 

Jamestown Meeting, 1907 Svo, 3 00 

BASHORE Outlines of Practical Sanitation 12mo, *1 25 

Sanitation of a Country House 12mo, 1 00 

Sanitation of Recreation Camps and Parks 12mo, 1 00 

FOLWELL Sewerage. (Designing, Construction, and Maintenance.).. .Svo, 3 00 

Water-supply Engineering Svo, 4 00 

FOWLER Sewage Works Analyses 12mo, 2 00 

FUERTES Water-filtration Works 12mo, 2 50 

GERHARD Guide to Sanitary Inspections 12mo, 1 50 

Modern Baths and Bath Houses Svo, *3 00 

Sanitation of Public Buildings 12mo, 1 50 

The Water Supply, Sewerage, and Plumbing of Modern City Buildings. 

Svo, *4 00' 
HAZEN Clean Water and How to Get It Small Svo, 1 50 

Filtration of Public Water-supplies Svo, 3 00 

KINXICUTT, WINSLOW and PRATT Sewage Disposal Svo, *3 00 

LEACH Inspection and Analysis of Food with Special' Reference to State 

Control Svo, 7 50 

MASON Examination of Water. (Chemical and Bacteriological.). ... 12mo, 1 25 

Water-supply. (Considered principally from a Sanitary Standpoint.) 

Svo, 4 00 

MERRIMAN Elements of Sanitary Engineering Svo, *2 00 

OGDEN Sewer Construction Svo, 3 00 

Sewer Design 12mo, 2 00 

OGDEN and CLEVELAND Practical Methods of Sewage Disposal for Res- 
idences, Hotels and Institutions Svo, *1 50 

PARSONS Disposal of Municipal Refuse Svo, 2 00 


PRESCOTT and WINSLOW Elements of Water Bacteriology, with Special 

Reference to Sanitary Water Analysis 12mo, $1 50 

PRICE Handbook on Sanitation 12mo, * 50 

RICHARDS Conservation by Sanitation 8vo, 50 

Cost of Cleanness 12mo, 00 

Cost of Food. A Study in Dietaries 12mo, 00 

Cost of Living as Modified by Sanitary Science 12mo, 00 

Cost of Shelter 12mo, 00 

Laboratory Notes on Industrial Water Analysis 8vo, *0 50 

RICHARDS and WOODMAN Air, Water, and Food from a Sanitary Stand- 
point 8vo, 2 00 

RICKEY Plumbers', Steam-fitters', and Tinners' Edition (Building Mechan- 
ics' Ready Reference Series) 16mo, mor., *1 50 

RIDEAL Disinfection and the Preservation of Food 8vo, 4 00 

SOPER Air and Ventilation of Subways 12mo, 2 50 

TURNEAURE and RUSSELL Public Water-supplies 8vo, 5 00 

VENABLE Garbage Crematories in America 8vo, 2 00 

Method and Devices for Bacterial Treatment of Sewage 8vo, 3 00 

WARD and WHIPPLE Freshwater Biology (In Press.) 

WHIPPLE Microscopy of Drinking-water 8vo, 3 50 

Typhoid Fever Small 8vo, *3 00 

Value of Pure Water Small 8vo, 1 00 


BuRT Railway Station Service 12mo, *2 00 

CHAPIN How to Enamel 12mo, *1 00 

EMMONS Geological Guide-book of the Rocky Mountain Excursion of the 

International Congress of Geologists Large 8vo, 1 50 

FERREL Popular Treatise on the Winds 8vo, 4 00 

FITZGERALD Boston Machinist 18mo, 1 00 

FRITZ Autobiography of John 8vo, *2 00 

GANNETT Statistical Abstract of the World 24mo, 75 

GREEN Elementary Hebrew Grammar 12mo, 1 25 

HAINES American Railway Management 12mo, 2 50 

HANAUSEK The Microscopy of Technical Products. (WiNTON.) 8vo, 5 00 

JACOBS Betterment Briefs. A Collection of Published Papers on Organ- 
ized Industrial Efficiency 8vo, 3 50 

METCALFE Cost of Manufactures, and the Administration of Workshops. 

8vo, 5 00 

PARKHURST Applied Methods of Scientific Management 8vo, *2 00 

PUTNAM Nautical Charts 8vo, 2 00 

RICKETTS History of Rensselaer Polytechnic Institute, 1824-1894. 

Small 8vo, 3 00 
ROTCH and PALMER Charts of the Atmosphere for Aeronauts and Aviators. 

Oblong 4to, *2 00 

ROTHERHAM Emphasised New Testament Large 8vo, 2 00 

RUST Ex-Meridian Altitude, Azimuth and Star-finding Tables 8vo, 5 00 

STANDAGE Decoration of Wood, Glass, Metal, etc 12mo, 2 00 

WESTERMAIER Compendium of General Botany. (SCHNEIDER.) 8vo, 2 00 

WINSLOW Elements of Applied Microscopy 12mo, 1 50